Inner-shell excitation and site specific fragmentation of poly

Inner-shell excitation and site specific fragmentation of poly(methylmethacrylate) thin film

Marcia C. K. Tinone and Kenichiro Tanaka Department of Synchrotron Radiation, The Graduate University for Advanced Studies, Tsukuba, 305 Japan and Photon Factory, National Laboratory for High Energy Physics, Tsukuba, 305 Japan

Junya Maruyama and Nobuo Ueno Department of Materials Science, Chiba University, Inage, Chiba, 263 Japan

Motoyasu lmamura and Nobuyuki Matsubayashi National Institute of Materials and Chemical Research, 1-I Higashi, Tsukuba, 305 Japan

(Received 12 October 1993; accepted 5 January 1994)

Soft x-ray excitations in the 250-600 eV photon energy range on poly(methylmethacrylate) (PMMA) result in ionic fragmentation of the original polymer with the most intense ions corresponding to CH: , H+, CH,’ , CHf, CHO+, and COOCH: . The photon energy dependence of ion desorption from thin films of PMMA was measured to investigate the primary steps in radiation induced decomposition following carbon and oxygen 1s electron excitations using monochromatic pulsed-synchrotron radiation. It was clearly found that the decomposition depends on the nature of the electronic states created in the excited species. The fragmentation pattern changes depending on the transitions of the 1s electron to a Rydberg orbital, an unoccupied molecular orbital or the ionization continuum. Moreover, the fragmentation occurs specifically around the site of the atom where the optical excitation takes place. Excitations from carbon and oxygen 1s to o* states seem to be specially efficient for ion production as observed in the case of CH: , CH,f , and CHf at 288.7 and 535.6 eV, and in the case of CHOf at 539.3 eV

I. INTRODUCTION

Inner-shell spectroscopies, mainly, x-ray photoelectron spectroscopy (XPS), inner-shell electron energy loss spectroscopy (ISEELS) (Ref. 1) and near-edge x-ray absorption fine structure (NEXAFS),’ have been successfully applied to the characterization of the electronic structure of small molecules in the gas phase or adsorbed on solid surfaces and, only recently, to large molecules and polymers.3 Inner-shell excitation using soft x rays is highly localized on specific atoms in molecules. The differences in core electron binding energies for different atoms are sufficiently large to allow selective excitation of different atoms or even the same atom at inequivalent chemical environments in the molecule. Moreover, depending on the atomic site of the core hole and the electronic configuration of the excited state, the following decay of the different final states created by photoexcitation can result in different chemical reactions. These states include configurations with the core electron excited to an antibonding valence orbital, a Rydberg orbital, a shape resonance, or the ionization continuum. Differently from excitations of valence electrons, excitations of core electrons are followed by the Auger decay,4 and the subsequent autoionization can leave the molecule in a multiple electron excited state that usually results in the molecular fragmentation.

Since core electrons are highly localized, there is an in- triguing possibility that the electronic relaxation and bond scission will also be localized around the atomic site of excitation. Questioning arises from the possibility of site- selective decomposition in spite of the large amount of energy remaining in the molecular system and from the possibility of losing the site specific character of the fragmentation by the subsequent core hole energy redistribution.

Evidence for such atom selective photochemistry has been found in studies of the ionic fragmentation of simple gaseous molecules. Eberhardt et al. have reported such site specificity from the observation of a strong resonance in the C+ and Oc yield spectra for the excitation of a carbon 1s electron to the T* molecular orbital of the CO group in acetone.5 Differ- ences in the relative yields of CF: have also been observed depending on the site of carbon or fluorine atoms 1s electron excitation in the CF,CH, molecule.6 Another good example was reported on the distinction between the central and ter- minal nitrogen and oxygen atom excitations for N,O molecules.7 The explanation for this selectivity, proposed for covalently bonded systems, involves the concept of two hole localization. It appears that the two valence holes created by Auger decay can be localized around a single atom for a time sufficiently long to allow bond rupture and ion desorption to occur.’

Despite various examples of site selective chemistry in the gas phase, the low ion desorption rates from solid molecules have hindered the investigation of the photochemistry of core excited states in the solid phase. Surface photochemistry is still at an early development stage and only a few papers have been devoted to this area of research. An early example has been reported by Hanson et al. on the desorption of ions following excitation by monochromatic soft x rays in the region of the carbon K-edge occurring in solid polymer thin films of polystyrene and polyethylene.g It has been shown that although a large amount of energy is depos- ited by a single photon, the subsequent relaxation and chemistry depend upon the nature of the initial excited state. The excitation to antibonding rr orbitals of the benzene ring in

5988 J. Chem. Phys. 100 (E), 15 April 1994 0021-9606/94/l 00(8)/5988/6/$6.00 0 1994 American Institute of Physics Downloaded 05 Dec 2002 to 129.49.56.78. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/jcpo/jcpcr.jsp

polystyrene does not result in significant bond rupture and ion desorption.

This paper is an extension of the photoion and photoelectron yields measurements of poly(methylmethacrylate) (PMMA) using synchrotron radiation.“‘” PMMA has been used as a positive photoresist for a long time because of its extremely high resolution and reasonable processing characteristics.12 Although its photochemical decomposition by ionizing radiation has been widely used in x-ray and electron beam lithography, the knowledge on the decomposition mechanism is very limited.13 The possibility that specific bonds can be formed or broken by the activation of selected atoms motivated the present research, following the discov- ery of a prominent resonant peak for CH: ions in prelimi- nary experiments near the carbon K-edge using low resolution monochromatic synchrotron radiation.”

In this paper, we report the first systematic detection of the main ions produced by core electron excitation in both carbon and oxygen K-edge regions for PMMA. The measurement of fragment ions was used as a direct probe of the radiation induced decomposition process. We show that site- specific excitation of a 1s electron of carbon or oxygen atoms at inequivalent sites results in different fragmentation patterns for PMMA. It was clearly found that the decomposition depends on the nature of the electronic states created in the excited species. In particular, the remarkable differences of oxygen excitation from different environments have been directly reflected in the yield spectra for different ions. The excitation site and antibonding orbitals involved in the photoexcitation could be clearly distinguished.

11. EXPERIMENT

The experiments were performed at the soft x-ray beamline BL13C at the Photon Factory of The National Labora- tory for High Energy Physics. This beamline14 was designed based on a cylindrical element monochromator (CEM) concept15 using undulator radiation from a 27-pole multipole wiggler/undulator. The photon energy range of the beam line is 200-1000 eV which covers the core-level excitation energy of carbon, nitrogen, and oxygen. A schematic diagram of the experimental setup is shown in Fig. 1. A time-of-flight mass spectrometer (TOF) was mounted in an ultrahigh vacuum (UHV) chamberlh which consists of two ultrahigh vacuum chambers, a chemical reaction cell, and a multi- sample introduction port. The TOF spectrometer is located at normal direction from the sample surface at an angle of 54.7” (magic angle) with respect to the incident photon beam. The soft x-ray beam was incident on the sample through a photon-flux monitor of Cu grid. The incident flux spectrum was recorded simultaneously as the photocurrent at the Cu grid. All spectra were normalized by this flux spectrum to correct for fluctuations in beam intensity. For the experiments reported here, a slit width of 100 pm was used for both entrance and exit slits, giving an energy resolution of 0.5 eV at 400 eV estimated from the XPS spectrum of MO&. The energy scale was calibrated using the two major x-ray absorption peaks of the carbon contaminants on the surfaces of optical elements of the monochromator at 284.7 and 291.0

Tinone et a/,: Fragmentation of thin film 5989

FIG. 1. Schematic diagram of the experimental setup. CEM, cylindrical element monochromator; AMP, amplifier; CFD, constant-fraction discrimi- nator; TAC, time-to-amplitude converter; ADC, analog-digital converter; MCA, multichannel analyzer.

eV.” The same calibration factor used at the carbon edge was used to calibrate the oxygen edge energy scale.

Both the total electron yield absorption spectrum (TEY) and partial ion yield spectra (PIY) were measured using the TOF mass spectrometer using positive and negative accelerating potentials, respectively. The TOF spectrometer consists of an accelerating plate, a drift tube of 9 cm and a multichannel plate detector (MCP). During single bunch operation of the Photon Factory storage ring, a soft x-ray pulse with a period of 624 ns and a width of 100 ps was incident to the sample. A l/312 divider is used to take the SR pulse timing from the 500 MHz microwave cavity frequency. In the TOF normal operation, the time-of-flight of the ions is measured as the time between the SR pulse and the amplified MCP ion detection pulse using an EG&G Ortec 567 time to amplitude converter. The total spectra is accumulated and read by an Laboratory Equipments ADC4803A analog to digital converter and LN-6400 multichannel analyzer. A typical TOF mass spectrum using 393 eV photon energy, -2.8 kV at the accelerating plate, -2.5 kV at the TOF drift tube, and -2.5 kV at the MCP is shown in Fig. 2. Two SR pulses can be seen at 624 ns interval. Due to the very good time reproduc- ibility of the SR pulses, heavy ions which have flight times longer than 624 ns can also be measured. For these heavier ions the time-of-flight is given by its position in the spectrum plus an integer multiple of the SR pulse interval (624 ns) corresponding to the number of cycles that have passed from the production of the ion to its detection. The mass assign-

J. Chem. Phys., Vol. 100, No. 8, 15 April 1994 Downloaded 05 Dec 2002 to 129.49.56.78. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/jcpo/jcpcr.jsp

5990 Tinone et al.: Fragmentation of thin film

600 . . . . I . . . . 1 ,* . . I ’ . I . . * . I . , . . t ’ . * * t . ’ . a* H’

- 1 ‘2

‘: * 400

60 70 80 -60 ' - : 30 3s 40 4s so 55 * .

10 IS -z) 25) : t -

0 1 2 j45h m/e

I.‘..,....,....,....,....,....,...., 0 200 400 600

Time-of-Flight /ns

FIG. 2. Typical ion time-of-flight spectrum obtained at 393 eV. The most intense Hf, CH+, CH;, CH:, CHO+, and COOCH: ions are indicated. The mass scale is given for increasing time cycles.

ments can be easily done by taking comparative spectra at different accelerating potentials. The most intense CH: , H+, CH: , CH+, CHO+, and COOCHZ ion peaks are indicated in Fig. 2. SR energy dependent partial ion yield spectra (PIY) can also be measured windowing the channels corresponding to the ion to be analyzed while scanning the photon energy. Better counting rates can be obtained using the TOF in re- verse mode, i.e., measuring the time-of-flight of the ions starting at the MCP ion detection pulse and stopping at the SR pulse.

Thin films of PMMA were prepared by spin-casting from methyl isobutyl ketone solution on an Au-evaporated Si(100) wafer. The thickness of the PMMA films was estimated to be inferior to 100 A as measured by a DEKTAK 3030 surface profiler. It is often assumed that for very thin insulating films on a conducting substrate, charging phenom- ena are entirely absent. In this work the charging up effect was considered negligible and there was no attempt for charge compensation. Since no significant alteration of the mass spectra was observed even after long x-ray exposures, 8 h, the x-ray damage effect was judged insignificant. The pressure in the measurement chamber was 3 X low7 Pa during the measurements, all performed at room temperature.

Ill. RESULTS AND DISCUSSION

A. Carbon K-shell excitation spectra

The structure of the PMMA monomer unit is shown in Fig. 3. Carbon and oxygen atoms at different sites are num- bered according to decreasing binding energies. Since the absolute values of ionization potentials (IP) measured by x-ray photoelectron (XPS) studies for thin films of PMMA

CH3 I

,/CC /-A,

\ c&o . . . ..-*

O;H, ! monomer unit --In

C-

FIG. 3. Poly(methylmethacrylate) monomer unit and numbering atoms at different chemical sites.

are subjected to large fluctuations within different experimental conditions, the IP were estimated as follows: The IP of the carbonyl carbon was aligned to the value of the equivalent carbon for gaseous methyl acetate (CH&OOCH3).r8 The IP’s for different carbon positions where estimated by the results obtained by XPS of PMMA.t9 Thus, we estimated the carbon IP’s as 294.85, 292.65, and 291.10 eV for C’, C2, and C3,4,5, respectively. There has been great controversy about the distinction of C3, C?, and C5 as discussed by XPS studies.” Nevertheless, throughout this work, we have neglected this difference in binding energies and have considered them equivalent.

In Fig. 4, the total electron yield spectrum of PMMA is compared to the total ion yield and partial ion yield spectra of the main ions Hf, CH+, CH,f , CH:, CHO+, and COOCH: . The hatched lines marked C’, C2, and C3*4*5 indicate the estimated ionization potentials. Six absorption features can be observed. The energies and proposed assignments are listed in Table I.

Outka and Stohr have shown that, using the building block approach, the C Is NEXAFS of PMMA can be pre- dicted by the ISEELS of formic acid and methyl formate used as model molecules.‘r Three peaks had been considered characteristic of the carboxylate group (-CO,-). (i) The first is the sharp C==O rr* resonance at approximately 289 eV, (ii) the second is the C-O cr* resonance at 296 ev; and (iii) the third is the C==O o* resonance at 302 eV The main differences between the gas-phase model molecules and the polymer were considered as a result of the contribution of the hydrocarbon backbone structures making the carboxylate features less prominent. In a similar way, in the following discussion the peaks assignments have been done by comparison to ISEELS and NEXAFS of formic acid, propanoic acid, methyl formate, and 2,2-dimethylmethylpropionate used as model molecules.

There are common features in all spectra. The first strong peak at 288.7 eV (feature 2) was first considered as due mainly to the C’ ls(C==O)-+rr*(C=O) transition as reported earlier in electron energy loss studies (EELS) of carboxylic acids.22 There is a shoulder at 287.5 eV which is assigned to C?,5 1s transitions to a CT* state mainly localized in the main chain carbons.23 The shoulder at 289.7 eV is a mixture of C ls+a*(C-H) from C2, and C?,5. The broad maxima (features 4, 5, and 6) at 292.3, 296.3, and 303.2 eV


Tinone et al.: Fragmentation of thin film 5991

in the continuum are attributed to the (+*(C-C, C-H),24 u*(C-0), and (+*(C=O) resonances.

Differently from the results of acetone by Eberhardt et al.,5 not only the ions produced by direct fragmentation of the CT==0 group give rise to the excitation observed at 288.7 eV, but all ions show a peak at this energy. Ritsko et al.23 have compared electron energy loss and photoemission spectra of PMMA to calculated results from all-valence-electron CNDO/S molecular orbital calculations of 2,2- dimethylmethylpropionate (DMMP). From these calculations, the peak at 288.7 eV can be assigned to several nearly degenerate components. The strongest consists of a transition from the carbonyl carbon, C’, to the first empty molecular orbital r*(G=O). The other significant contributions consist of transitions from C3 to the third, fourth, and fifth empty molecular orbitals mainly localized in the t;butyl as well as from C* to the fourth empty state mainly localized in the methoxy methyl group. The K-shell excitation spectra of CH,f and CH+ differ from that of other fragments chiefly by a prominent peak at 288.7 eV. For these ions this peak is consistent with a transition of C3 ls+c~*(C-C) localized at the main chain and C* ls+c~*(O-C) at the methoxy group. This prominent peak at 288.7 eV especially in the curves which correspond to the methyl group fragmentation, can be taken as a strong indication of the higher efficiency of ion production by the (T* excitation. Similar results were observed in polystyrene films which shows that T* transitions of the benzene ring do not result in bond rupture and ion desorption.’

There is an overall agreement of the present results and the C 1s NEXAFS of condensed propanoic acid,= demon- strating once again the viability of using molecule analogs to assist the analysis of the spectra of polymers.26 Within the building block model the C 1s spectra of PMMA is approximately the sum of the methyl formate and the propiolic acid spectra with additional detail arising from chemical shifts among the chemically inequivalent carbons. This approxima- tion gives some insight into spectra1 assignments but it does

280 290 300 310

Photon Energy /eV

FIG. 4. Total electron yield (TEY), total ion yield (TIY), and partial ion yields of H+, CH+, CH;, CH; , CHO+, and COOCH: near carbon K-edge. Solid curve represents the result of curve fittings. Hatched lines C’, C’, and C3.” indicate the estimated ionization potentials described in the text.

TABLE I. Photon energies, term values (IP-E) and proposed assignments for the features observed in the C K-shell excitation spectra of PMMA.

Term value (eV) Assignment

CL0 0 C’H3 CL0 0 (-3 Energy I I I I

Feature W 0 C’H, C3-C5H2 0 C’H, C3-CsH2

1 287.5 .*. . . . 3.6 . . . . . . u*(c-C) 2 288.7 6.2 4.0 2.5 Tr*(c=o) u*(o-cH3) a*(C-C) 3 289.7 ‘.’ 2.9 1.4 . . . a*(C-H)

IP(cBs)a o*(C-H)

291.1 ... . . . . . . . . . IP . . . 4 292.3 2.5 . . . 71.2 d(C-H) ... o*(c-C)

IP(C2) 292.65 ... . . . . . . . . . IP . . . IP(C’)’ 294.85 ... . . . . . . P . . . ..*

5 296.3 -1.4 . . . -5.2 o*(C-OCH,) ... u*(c=o) 6 303.2 -8.3 . . . . . . u*(cbo) ... . . .

“Estimated IP’s, see text.


not account for interactions within nearby side chain carbonyl groups and long-range interactions between monomer units side chains. The spin cast PMMA used during this studies is a mixture of isotatic, heterotatic, and syndiotatic (4.1:36.9:59.0) PMMA and is considered to be randomly dis- tributed on the surface. In order to investigate long range interactions between side chains, further experiments using polarized light, oriented and tacticity pure PMMA films are planned for a near future.

8--

B. Oxygen K-shell excitation spectra

In a similar way to the carbon K-edge, the oxygen K-edge IP’s were estimated as follows: The IP of the methoxy oxygen was aligned to the value of the equivalent oxygen of gaseous methyl propionate (CH,CH,COOCH,).27 The IP for the carbonyl oxygen was estimated by the results obtained by XPS of PMMA.19 Thus, we estimated the oxygen IP’s as 539.32 and 537.82 eV for 0’ and O*, respectively. In Fig. 5 the total electron yield spectrum of PMMA is compared to the total ion yield and partial ion yield spectra of the main ions H”, CH+, CH:, CH:, CHO+, and COOCI-I; .

7 d .

7 “! 8-- UI . E . 2 . Q .

2 a . 2 I .

4 -- s z

The hatched lines marked 0’, and O2 indicate the estimated m .

ionization potentials. Six absorption features can be ob- cz .

served. The energies and proposed assignments are listed in Table II.

FIG. 5. Total electron yield (TEY), total ion yield (TIY), and partial ion yields of H+, CH+, CH; , CH; , CHO+, and COOCH; near oxygen K-edge. Solid curve represents the result of curve fittings. Hatched lines 0’, and 0’ indicate the estimated ionization potentials described in the text.

There are six common features in all spectra. The first strong peak at 531.5 eV is due to the 0 1s to rr* (c---O) transition as reported earlier in ISEELS of methyl formate and propanoic acid.** The second peak at 534.3 eV can be assigned to a mixture of 0 ls(OCH,)+n*(C---=O) charge transfer transition and 0 ~~(C=O)~(T*(C~-O~). Peak 3 at 535.6 eV can be assigned to 0 ls(OCH&+a*(C*-0’) resonance. The higher energy features 4, 5, and 6 are attributed to unresolved 0 ls(OCH&+Rydberg/ cr*(C-OCHJ, 0 ls(CLO)-+a*(C=O) and shake up ~((1F---=o)--+rr*(C=O) transitions, respectively.‘* In the PMMA 0 K-edge there is also an overall agreement with the NEXAFS main features of condensed propanoic acid;‘* the c----O rr* resonance at 532.8 eV, the C-O (+* resonance at approximately 540.1 eV, and C==O u* resonance at 544.4 eV are consistent with species containing a carboxyl group. The clear qualitative differences found among the different fragments can be related in a simple manner to the localized transitions which originate the fragmentation. The third feature is notably prominent in the spectra of CHf, CH:, and CH:. The efficient production of these ions can be understood as a result of the C-O bond breaking of the methoxy group as a result of the 0 ls(OCH&+cr*(C*-0’) transition. The dominant feature in the CHOf spectrum at 539.3 eV is considered as a result of the 0 ls(OCH&+a*(C’-0’) transition. These interesting differences observed among the spectra of the fragments indicate that the fragmentation process is very sensitive to the unoccupied electronic structure in the immediate vicinity of the oxygen atoms.

TABLE II. Photon energies, term values (IP-E) and proposed assignments for the features observed in the 0 K-shell excitation spectra of PMMA.

Term value (eV) Assignment Energy

Feature (eV) o’-CH, c--o2 O’-CH, c==02

C. Site specific fragmentation

Electron yield and more recently, fluorescence yield have been the most frequently used detection techniques in

1 531.5 ... 6.3 2 534.3 5.0 3.5 3 535.6 3.7 2.2

IP(02)8 537.82 ... ... 4 539.3 0.02 ...

IP(O’)B 539.32 **. . . .

5 546.1 ... -8.3 6 552.0 ..* - 14.2

aEstimated IP’s, see text.

5992 Tinone et al.: Fragmentation of thin film

TIYH5

530 540 550 5$0 Photon Energy /eV

. . . Tr*(c==o) n*(cbo) (r*(C-OCH&

u*(O-CHJ Rydberg . . . IP

Rydberg/a*(C-OCHJ ... IP *.. . . . a* (c=o) . . . Shake up


40

30

20

10

0

290 300

Photon Energy /eV

FIG. 6. Experimental points and curve fitted for the CH: partial ion yield near carbon K-edge. The observed features are fitted with six Gaussian curves and one step function at 292.13 eV.

NEXAFS. The decay process after the primary core excitation, however, can be more directly investigated by ion yield detection. The ion desorption can be considered as a tertiary step after the primary excitation and the Auger decay.4 Only those ions which have sufficient energy to overcome the surface potential barrier and survive reneutralization will escape into vacuum. This explains the small desorption rates in the ion yield spectra which have hindered the utilization of the ion yield detection technique for NEXAFS.29 Surprisingly, we have obtained for PMMA, not only total ion yield but also PIY spectra with reasonable intensities and resolution which still reveal the localized character of the primary excitation. Comparing the PIY spectra to the TE?Y spectrum (Figs. 4 and S), one can observe that even with low intensi-

*O? 18--

16--

14--

12 --

Tinone et a/.: Fragmentation of thin film 5993

TABLE III. Photon energies and branching ratios for the features observed in the C Is spectra of PMMA.

hv TEY TIY H+ CH’ CH; CH; CHO+ COOCH; Feature (eV) (k) (I) (%) (%) (%) (%) (%) (o/o)

1 287.5 5.5 2.6 3.7 0.9 2.1 2.3 2.4 2.3 2 288.7 22.3 33.4 29.8 62.7 58.5 29.3 19.6 21.7 3 289.7 7.2 11.0 8.3 5.8 5.1 8.7 4.5 4.0 4 292.3 6.2 9.0 5.1 3.9 1.6 8.6 8.3 12.5 5 296.3 22.4 18.1 19.4 3.2 7.0 23.7 28.5 22.9 6 303.2 36.4 25.9 33.7 23.5 25.7 27.4 36.7 36.6

ties, the features in the first are better defined and more sensitive to photon energy than in the total ion yield or total electron yield spectra. In order to make a more quantitative comparison among different fragment ions, the experimental data were fitted to six Gaussian peaks under each observed feature and one step function representing the onset of the ionization continuum. The experimental data and fitted curves for the CH: ion near the carbon and oxygen edges are given as example in Figs. 6 and 7. The branching ratios within different features are given in Tables III and IV for carbon and oxygen edges, respectively. These branching ratios are given as the area under each fitted Gaussian peak normalized to the total area under the curve less the onset of the ionization continuum. More efficient production of different ions at a specific transition is easily observed by the increase in the percentage of the corresponding feature area.

While the TEY spectrum of PMMA reflects the total x-ray absorption cross section, the TIY reflects the desorption efficiency of the most abundant ions (CH: and H+ near carbon K-edge and CH:, H+, and CHOf near oxygen K-edge). The PIY spectra near carbon edge are all very similar, with exceptions made to CH+ and CH: . In Table III we can observe that there is a drastic change in the branching ratios of CH+ and CH: as a result of the increased contribution of feature 2. As has been observed before, this strong peak was assigned to transitions localized mainly in the main chain C-C bondings and in the methoxy group. These experiments alone can not distinguish the origin of these CH+ and CH: ions from the main chain or from the methoxy group fragmentation. The branching ratios of H+, CH:, COOCH;, and TEY are very similar and, as will be discussed below, are in agreement with the idea of separating the measured ions as those which are produced by a single

TABLE IV. Photon energies, and branching ratios for the features observed in the 0 1s spectra of PMMA.

10. 530 540 550 560

Photon Energy IeV

FIG. 7. Experimental points and curve fitted for the CH: partial ion yield near oxygen K-edge. The observed features are fitted with six Gaussian curves and one step function at 538.57 eV.

+ CH+ CH+ CH+ CHO+ COOCH; Feature (t;) Ey k: (&) (%) (%; (%i (%) (o/o)

1 531.5 10.0 13.1 14.2 8.2 7.8 15.0 3.6 11.1 2 534.3 5.2 5.0 8.5 3.2 2.2 4.2 4.4 6.3 3 535.6 6.7 14.1 4.7 13.2 23.9 22.3 0.6 5.3 4 539.3 36.5 43.6 41.5 27.3 30.4 38.8 59.3 33.4 5 546.1 21.2 14.0 16.4 20.5 14.8 11.5 24.0 26.0 6 552.0 20.4 10.2 14.7 27.6 20.9 8.2 8.1 17.9


bond breaking and those that need more than one bond breaking to be produced.

Differently from the carbon K-shell excitation, the PIY spectra of oxygen K-shell excitation are remarkably different from the TEY spectrum. The CHC, CH:, and CH: spectra are characterized by a stronger peak at 535.6 eV assigned to the 0’ ls(OCH,)+a*(C2-0’) also noticed by the increased percentage of its branching ratio in Table IV. The CHOf spectrum is dominated by the peak at 539.3 eV (feature 4) that represents 59.3% of the transitions that result in CHO+ desorption. Feature 2 is considerably less intense than feature 1 in all spectra except for CHOf ions. This observation can be easily understood as an effect of the charge transfer from the methoxy group to the carbonyl group that facilitates the production of CHO+ ions, but has small contribution for the production of other fragments. It is also interesting to note the difference between features 2 and 3. Even though both are assigned to transitions from the same atomic site, methoxy group oxygen 1s electrons, there is a large difference in efficiency for producing CH+, CH:, and CH: or CHO+. This is an indication of the importance of the ,unoccupied orbital characteristics. Few fragment ions are expected to be produced by transitions to nonbonding unoccupied orbitals. However, localized bond scissions can be enhanced by transitions to unoccupied orbitals with a strong antibonding character at the specified bonding.

The core excitation and ionization of gaseous HBr, CHsBr, HI, and CHsI, was shown to be a two step relaxation process involving a fast neutral dissociation followed by the autoionization of the excited fragments.30 For heavier fragments such as those measured in PMMA this phenomenon is considered to disappear. Due to strong interactions in the solid surface, only a minor part of the excited states produced in the photoexcitation step effectively result in ion desorption. Figure 8 schematically shows the ion desorption mechanism proposed for PMMA. After the first excitation, highly charged molecules are produced during the Auger process. The following energy decay proceed through different paths. In the fast and energetic path the highly charged molecules easily fragment. The energy randomization in the molecule is a competing path which results in the production of less energetic fragments or, as noted before, rapidly deac- tivate before fragmentation. The observed spectra of PMMA can be explained by a balance between the two main fragmentation paths; the more energetic path represented with a solid line in Fig. 8 and the less energetic path that occurs after partial energy randomization in the excited molecule, represented by the broken line. This mechanism agrees with the observation that the most remarkable differences between the TEY spectrum and partial ion yield spectra occur with CH+, CH:, and CHOC ions. The energetic path is more favorable to produce these ions noting that not only a single bond breakage, but also the ejection of H or H, is necessary. Since the energetic path also represents the fast nonenergy- randomized process, it is not unexpectedly that the spectra measured for these ions would also be those which more clearly show site selectivity. The similarity within the partial ion spectra of Hf, CH:, and COOCH: and the TEY spectrum can be analogously understood by the easily production

EXCITATION

5994 Tinone et a/.: Fragmentation of thin film

1 E&y I- +pyzfq

: :

AUGER i-r’ t

: v : ION FRAGMENTATION

: I

FIG. 8. Proposed mechanism of ion fragmentation and desorption of PMMA by core electron excitation. ‘Ike fast and energetic path is represented by solid lines and the less energetic path after partial energy randomization is represented by dashed lines. The site specific characteristic of the partial ion yield spectra is enhanced for those fragments for which the energetic path is favorable.

of these ions by a single bond scission which can occur even by less localized excitations and after partial energy randomization. In the gas phase, the reneutralization path is less probable and all produced ions, either from fast or slow reactions are collected by ion detection, the resultant spectrum being an average of all decomposition paths. The observation of site specific ion desorption in PMMA is then a result not only of the existence of localized transitions and fragmenta- tions, but also a result of the selective measurement of those ions produced by the fast and energetic path which are directly related to the localized excitations. Although ion desorption rates are considerably smaller than electron emis- sion rates, ion detection for solid samples has proven to be a valuable detection technique to assist in peak assignments and to investigate surface reaction mechanisms.

IV. CONCLUSIONS

In the present work, interesting information about the photon stimulated ion fragmentation of thin films of poly (methylmethacrylate) (PMMA) has been obtained near carbon and oxygen K-edges. The most intense ions correspond to CH; and Hf ions. The CH+, CH: , CHO+, and COOCH; ions are also observed in considerable intensities. Using soft x rays, the ion desorption mechanism of PMMA was shown to be directly related to the excitation from 1s core electrons at specific atomic sites to unoccupied orbitals at neighboring bondings. Near the carbon K-edge the strong peak at 288.7 eV was assigned to a mixture of carbon 1s transitions to unoccupied molecular orbitals, mainly, a rr* state localized at the carbonyl group and (+* state localized at the main

J. Chem. Phys., Vol. 100, No. 8, 15 April 1994

Downloaded 05 Dec 2002 to 129.49.56.78. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/jcpo/jcpcr.jsp

Tinone et a/.: Fragmentation of thin film 5995

chain carbon. The efficient production of CH: and CHf at this energy, is taken as evidence of the localized contribution of the c* state at the main chain carbons. Near the oxygen K-edge, the strong peak localized at 537.7 eV can be understood as an efficient production of CH:, CH:, and CHf ions by the excitation of 0 1s electrons to (T* states localized at the methoxy group. The broader feature at 539.3 eV can be assigned to transitions from the methoxy 0 1s to the (+* state of the carbonyl-methoxy C-O bonding. This would explain the good efficiency of production of CHO+ ions observed at this energy.

This work has also demonstrated the viability of ion detection application as a powerful technique for investigation of core excitation. The partial ion yield detection is considered as a complementary technique, to normal electron yield NEXAFS. The partial ion yield detection has shown one main advantage over electron and ‘TIY detection, namely, more clear differentiation and assignment of transitions from different atomic sites by considering the conectivity of the fragments in the original molecule. The extension of this research to different molecules are planned especially to investigate the origin of CH: , CH: , and CHf ions. This work demonstrates the viability of using core excitation as a means to produce localized decomposition of the well known resist PMMA. These results can be directly applied for the con- struction of future molecular electronic devices31 in which the ability of controlling very localized reactions is one of the fundamental aspects.

ACKNOWLEDGMENTS

This work was performed under the approval of the Pho- ton Factory Program Advisory Committee (Proposal No. 92- 150). This research was financially supported by a Grant-In- Aid from the Ministry of Education, Science and Culture. M. C. K. T. would like to thank CNPq in Brazil for financial support.

‘A. F’. Hitchcock, J. Electron. Spectrosc. Relat. Phenom. 25, 245 (1982). ‘(a) J. StGhr and D. A. Outka, J. Vat. Sci. Technol. A 5, 919 (1987); (b) J.

Stijhr, NEXAFS Spectroscopy, Vol. 25 in Springer Series in Surface Sci- ence (Springer, Berlin, Heidelberg, 1992).

3(a) G. Tourillon and Y. Jugnet, J. Chem. Phys. 89, 1905 (1988); (b) A. P. Hitchcock, G. Tourillon, R. Garret, G. P. Wil l iams, C. Mahatsekake, and C. Andrieu, J. Phys. Chem. 94, 2327 (1990); (c) T. Ohta, K. Seki, T. Yokoyama, 1. Morisada, and K. Edamatsu, Phys. Ser. 41, 150 (1990); (d) A. P. Hitchcock, S. G. Urquhart, and E. G. Rightor, J. Phys. Chem. 96, 8736 (1992).

‘(a) P. J. Feibelman and M. L. Knotek, Phys. Rev. 18, 6531 (1978); (b) M. L. Knotek and P. J. Feibelman, Surf. Sci. 90, 78 (1979); (c) A. P. Hitch- cock, Phys. Ser. T31, 159 (1990).

sW. Eberhardt, T. K. Sham, R. Can; S. Kmmmacher, M. Strongin, S. L. Weng, and D. Wesner, Phys. Rev. Lett. 50, 1038 (1983).

6K. Muller-Dethlefs, M. Sander, L. A. Chewter, and E. W. Schlag, J. Phys. Chem. 88, 6098 (1984).

7J. Murakami, M. C. Nelson, S. L. Anderson, and D. M. Hanson, J. Chem. Phys. 85, 5755 (1986).

‘(a) D. E. Ramaker, C. T. White, and J. S. Murday, J. Vat. Sci. Technol. 18, 748 (1981); (b) P. J. Feibelman, Surf. Sci. 102, 151 (1981); (c) D. E. Ramaker, C. T. White, and J. S. Murday, Phys. Len. 89A, 211 (1982).

‘D. M. Hanson, S. L. Anderson, M. C. Nelson, G. P. Wil l iams, and N. Lucas, J. Phys. Chem. 89, 2235 (1985). ’

‘ON Ueno, Y. Kobayashi, T. Sekiguchi, H. lkeura, K. Sugita, K. Honma, K. Tanaka, E. Orti, and R. Viruela, J. Appl. Phys. 72, 5423 (1992).

‘IN. Ueno, M. Komada, Y. Morimoto, M. C. K. Tinone, M. Kushida, K. Sugita, K. Honma, and K. Tanaka, Jpn. J. Appl. Phys. 32, 229 (1993).

‘*(a) N. Ueno and K. Sugita, Jpn. J. Appl. Phys. 25, 1455 (1986); (b) N. Ueno, Y. Doi, K. Sugita, S. Sasaki, and S. Nagata, J. Appl. Polym. Sci. 34, 1677 (1987); (c) R. F. Pease, Jpn. J. Appl. Phys. 31,4103 (1992); (d) G. D. Kubiak, E. M. Kneedler, R. Q. Hwang, M. T. Schulberg, K. W. Berger, J. E. Bjorkholm, and W. M. Mansfield, J. Vat. Sci. Technol. B 10, 2593 (1992); (e) K. Early, D. M. Tennant, D. Y. Jeon, P. P. Mulgrew, A. A. * MacDowell, and 0. R. Wood 11, ibid. 10, 2600 (1992).

13(a) M. Przybylski, M. Stamm, and R. Zietz, J. Phys. (Paris) 48, 1351 (1987); (b) N. Ueno, T. Mitsuhata, K. Sugita, and K. Tanaka, Jpn. J. Appl. Phys. 27, 1723 (1988); (c) in Polymers in Microlithography, ACS sympo- sium series No. 412, edited by E. Reichmanis, S. A. MacDonald, and T. lwayanagi (American Chemical Society, Washington, DC, 1989), Chap. 26, pp. 424-436; (d) B. W. Yates, D. M. Shinozaki, A. Kumar, and W. J. Meath, J. Polym. Sci. B 30, 185 (1992); (e) G. M. Wells, J. W. Taylor, E Certina, D. Pearson, and J. MacKay, J. Vat. Sci. Technol. B 10, 3252 (1992).

14N. Matsubayashi, H. Shimada, K. Tanaka, T. Sato, Y. Yoshimura, and A. Nishima, Rev. Sci. Instrum. 63, 1363 (1992).

“C. T. Chen and F. Sette, Rev. Sci. Instrum. 60, 2533 (1989). 16N. Matsubayashi, I. Kojima, M. Kurahashi, A. Nishima, A. ltoh, and T.

Utaka, Rev. Sci. lnstrum. 60, 2533 (1989). “D. A. Outka and J. Stiihr, J. Chem. Phys. 88, 3539 (1988). “S. R. Smith and T. D. Thomas, J. Am. Chem. Sot. 100, 5459 (1978). 19A Naves de Brito, M. P. Keane, N. Correia, S. Svenson, U. Gelius, and B.

J.‘Lindberg, Surf. Interface Anal. 17, 94 (1991). “A. Naves de Brito, N. Correia, S. Svensson, and H. Agren, J. Chem. Phys.

95, 2965 (1991). ‘ID. A. Outka and J. Stiihr, in Chemistry and Physics of Solid Surfaces VII,

Vol. 10 in Springer Series in Surface Science, edited by V. R. Vanselow and R. Howe (Springer, Berlin, Heidelberg, 1988), Chap. 6, pp. 201-210.

“1. lshii and A. P. Hitchcock, J. Electron Spectrosc. Relat. Phenom. 46, 55 (1988).

uJ. J. Ritsko, L. J. Brillson, R. W. Bigelow, and T. J. Fabish, J. Chem. Phys. 69, 3931 (1978).

%J. St&r, D. A. Outka, K. Baberschke, D. Arvanitis, and J. A. Horsley, Phys. Rev. B 36, 2976 (1987).

25D. A. Outka, J. Stiihr, R. J. Madix, H. H. Rotermund, B. Hermsmeier, and J. Solomon, Surf. Sci. 185, 53 (1987).

26A. P. Hitchcock, S. G. Urquhart, and E. G. Rightor, J. Phys. Chem. 98, 9736 (1992).

27B. E. Mills, R. L. Martin, and D. A. Shirley, J. Am. Chem. Sot. 98, 2380 (1976).

=B. Sjijgren, A. N. de Brito, N. Correia, S. Lunnell, B. Wannberg, U. Gelius, and S. Svensson, J. Electron Spectrosc. Relat. Phenom. 59, 161 (1992).

“R. McGrath, 1. T. McGovern, D. R. Warburton, G. Thornton, and D. Nor- man, Surf. Sci. 178, 101 (1986).

“Op. Morin and 1. Nenner, Phys. Rev. Lett. 56, 1913 (1986). 31E. Sigmund, P. Gribi, and G. lsemann, Appl. Surf. Sci. 65/66, 342 (1993).


Documents

Inner-shell excitation and site specific fragmentation of poly