Bond insertion, complexation, and penetration pathways of vapor-deposited aluminum atoms with HO- and CH3O-terminated organic monolayers

Published on Web 04/20/2002

Bond Insertion, Complexation, and Penetration Pathways of

Vapor-Deposited Aluminum Atoms with HO- and

CH₃O-Terminated Organic Monolayers

Gregory L. Fisher,^§Amy V. Walker,^†Andrew E. Hooper,^|Timothy B. Tighe,^†

Kevin B. Bahnck, Hope T. Skriba,^†Michael D. Reinard,^†Brendan C. Haynie,^†

Robert L. Opila,*^,‡Nicholas Winograd,*^,†and David L. Allara*^,†

Contribution from the Department of Chemistry, The PennsylVania State UniVersity,

UniVersity Park, PennsylVania 16802, and Department of Materials Science,

UniVersity of Delaware, Newark, Delaware 19617

Received October 12, 2001

Abstract: The interaction of vapor-deposited Al atoms with self-assembled monolayers (SAMs) of HS-

(CH₂)₁₆-X (X ) -OH and -OCH₃) chemisorbed at polycrystalline Au{111} surfaces was studied using

time-of-flight secondary-ion mass spectrometry, X-ray photoelectron spectroscopy, and infrared reflectance

spectroscopy. Whereas quantum chemical theory calculations show that Al insertion into the C-C, C-H,

C-O, and O-H bonds is favorable energetically, it is observed that deposited Al inserts only with the OH

SAM to form an -O-Al-H product. This reaction appears to cease prior to complete -OH consumption,

and is followed by formation of a few overlayers of a nonmetallic type of phase and finally deposition of a

metallic film. In contrast, for the OCH₃SAM, the deposited Al atoms partition along two parallel paths:

nucleation and growth of an overlayer metal film, and penetration through the OCH₃SAM to the monolayer/

Au interface region. By considering a previous observation that a CH₃terminal group favors penetration as

the dominant initial process, and using theory calculations of Al-molecule interaction energies, we suggest

that the competition between the penetration and overlayer film nucleation channels is regulated by small

differences in the Al-SAM terminal group interaction energies. These results demonstrate the highly subtle

effects of surface structure and composition on the nucleation and growth of metal films on organic surfaces

and point to a new perspective on organometallic and metal-solvent interactions.

1. Introduction

organic groups at the vacuum interface, recently has provided

an approach to overcoming this problem.^5-16Not only are SAMs

of use as generalized models for metal-organic interactions

involving a wide variety of organic surfaces, but they also have

become of direct interest recently with the discovery that

electronic devices can be fabricated using SAMs with deposited

metal contacts. The study of the metal-SAM interactions in

The underlying chemistry of the interaction of metal atoms

with organic thin films and polymer surfaces has important

implications in many areas of science and technology.¹Elucida-

tion of the fundamental reaction pathways has proven to be

particularly challenging over the years since for most organic

systems the precise nature of the organic surface is difficult to

ascertain. The use of self-assembled monolayers (SAMs),^2-4

which provide a known surface density of uniformly organized

(5) Jung, D. R.; Czanderna, A W. Crit. ReV. Solid State Mater. Sci. 1994, 19,

1-54.

(6) Jung, D. R.; Czanderna, A. W.; Herdt, G. C. J. Vac. Sci. Technol. 1996,

A14, 1779-1787.

* To whom correspondence should be addressed. E-mail (D.L.A.):

dla3@psu.edu.

(7) Czanderna, A. W.; King, D. E.; Spaulding, D. J. Vac. Sci. Technol. 1991,

A9, 2607-2613.

^†The Pennsylvania State University.

(8) Jung, D. R.; King, D. E.; Czanderna, A. W. J. Vac. Sci. Technol. 1993,

A11, 2382-2386.

^‡University of Delaware.

(9) Konstadinidis, K.; Zhang, P.; Opila, R. L.; Allara, D. L. Surf. Sci. 1995,

338, 300-312.

^§Present address: Los Alamos National Laboratories, Los Alamos, NM

87545.

(10) Tarlov, M. J. Langmuir 1995, 11, 80-89.

(11) Dressick, W. J.; Dulcey, C. S.; Georger, J. H., Jr.; Calabrese, G. S.; Calvert,

J. M. J. Electrochem. Soc. 1994, 141, 210-220.

(12) Jung, D. R.; King, D. E.; Czanderna, A. W. Appl. Surf. Sci. 1993, 70/71,

127-132.

^|Present address: Motorola Corporate Laboratories, Tempe, AZ 85284.

Present address: Pfizer Laboratories, Inc., Groton, CT 06340.

(1) For example, see: Mittal, K. L., Ed. Metallized Plastics 5 & 6: Funda-

mental and Applied Aspects; VSP International Science Publishers: The

Netherlands, 1998; and earlier volumes. Mittal, K. L., Ed. Metallized

Plastics: Fundamentals and Applications; Marcel Dekker: New York,

1998.

(13) Herdt, G. C.; Czanderna, A. W. Surf. Sci. Lett. 1993, 297, L109-L112.

(14) Hirose, Y.; Kahn, A.; Aristov, V.; Soukiassian, P. Appl. Phys. Lett. 1996,

68, 217-219.

(2) Ulman, A. Chem. ReV. 1996, 96, 1533-1554.

(15) Herdt, G. C.; King, D. E.; Czanderna, A. W. Z. Phys. Chem. 1997, 202,

163-196.

(3) Dubois, L. H.; Nuzzo, R. G. Annu. ReV. Phys. Chem. 1992, 43, 437-463.

(4) Swalen, J. D.; Allara, D. L.; Andrade, J. D.; Chandross, E. A.; Garoff, S.;

Israelachvili, J.; McCarthy, T. J.; Murray, R.; Pease, R.; Rabolt, J.; Wynne,

K.; Yu, H. Langmuir 1987, 3, 932-950.

(16) Mittal, K.; Lee, K. W., Eds. Polymer Surfaces and Interfaces: Charac-

terization, Modification, and Applications; VSP International Science

Publishers: The Netherlands, 1999; pp 189-224.

9

5528

J. AM. CHEM. SOC. 2002, 124, 5528-5541

Vapor-Deposited Aluminum Atoms

A R T I C L E S

these structures should prove very useful in designing ohmic

contacts¹⁷. Also, from a fundamental perspective, SAMs offer

an interesting and useful complement to studies of the solvation

and electron-transfer reactions of metal atoms with gas-phase

molecules and molecular clusters.^18-20While the fundamental

solvation and chemical interaction mechanisms of a given metal

atom-molecule (or molecular group) system are fixed (e.g.,

electrostatic, van der Waals, electron exchange), variation of

the system geometry can vary the degree of the interactions. In

particular, with SAMs the interactions are constrained to occur

within a quasi-2-D planar geometry, in contrast to the 3-D

configurations allowed in a single phase.

four additional Al atoms continue to undergo redox interactions

with the terminal group leading to an organo-aluminum complex

with nonzero Al valence states. In a subsequent study²⁷involving

a HO₂C-terminated alkanethiolate/Au{111} SAM, we reported

that Al atoms react with the acid group in an ∼1:1 average

stoichiometry with no penetration into the SAM. An intriguing

aspect of this study is that while the chemical degradation of

the CO₂H group occurs during the initial deposition of the Al

atoms, some 15-20% fail to react, even with continued

deposition up to many Al atoms per molecule.

Two main conclusions have emerged from these initial

studies. First, when a reactive group is present at the alkyl chain

terminus, nucleation and growth of a metallic film occur only

at the vacuum/film interface, and the first several layers of Al

atoms appear to form an organo-aluminum, dielectric layer prior

to the growth of a metallic film. Second, when the chain

terminus contains an unreactive CH₃group, penetration through

the SAM to the S/Au interface occurs. It was proposed that

this penetration occurs via a thermally activated lateral hopping

process of the SAM molecules that leads to the creation of

transient holes allowing transport of nearby Al atoms directly

to the S/Au interface.²⁶

Of the various metals to consider, Al has been of particular

interest to us because of its common use in metallization of

polymers and as top contacts in organic electronic devices.^16,21

Further, in terms of a complement to gas-phase solvation studies,

as noted above, there has been interest in characterizing the

interaction of Al atoms with clusters of small gas molecules

such as H₂O, NH₃, and (CH₃)₂O,^22-24closely analogous to the

interaction of Al atoms with terminal SAM groups such as

-NH₂, -OH, and -OCH₃.

Recently, we reported on the deposition of Al atoms onto

H₃CO₂C- (methyl ester) and H₃C-terminated alkanethiolate/

Au{111} SAMs.^25,26It was observed that Al atoms exhibit an

unexpectedly subtle and discriminating chemistry with these

monolayers. In both cases, no reaction of the Al occurs with

the -CH₂- groups of the alkyl chains. In the case of the -CH₃

termination, deposited Al atoms were observed to penetrate

through the monolayer to the S/Au interface where it appeared

that each Al atom inserted into a Au-S bond to form an

aluminum thiolate species. Upon completion of this adlayer,

Al was observed to begin depositing exclusively at the vacuum/

SAM interface. In the case of the H₃CO₂C-terminated SAM,

the deposited Al atoms do not penetrate through the SAM but

rather react in a 1:1 stoichiometry with the carbonyl portion of

the ester functionality while leaving the C-O (ether) linkage

intact. Further, past the first deposited Al atom per group, about

From these studies, the CdO carbonyl oxygen atoms within

the -CO₂CH₃and -CO₂H terminal functional groups clearly

are implicated as a critical reaction center in the metal-organic

interaction. In this paper, we examine this interaction in more

detail by studying the interactions of Al atoms with HO-

(hydroxy) and H₃CO- (methoxy) terminated alkanethiolates on

Au{111}. These functional groups can be viewed simplistically

as representing constituents of the -CO₂CH₃group. Studying

each constituent thus could be expected to shed light on the

subtle preferences of Al for specific molecular reaction sites in

this family of O-containing functional groups. In particular, a

detailed characterization of the interplay between Al chemical

interaction and penetration pathways is critical in establishing

a fundamental basis for understanding nucleation and growth

processes of Al films on a variety of O-containing organic

surfaces.

(17) For example, see: (a) Tour, J. M. Acc. Chem. Res. 2000, 33, 791-804.

(b) Reed, M. A.; Tour, J. M. Sci. Am. 2000, 282, 86-93. (c) Metzger, R.

M.; Xu, T.; Peterson, I. R. J. Phys. Chem. B 2001, 105, 7280-7290. (d)

Metzger, R. M.; Chen, B.; Hopfner, U.; Lakshmikantham, M. V.;

Vuillaume, D.; Kawai, T.; Wu, X. L.; Tachibana, H.; Hughes, T. V.;

Sakurai, H.; Baldwin, J. W.; Hosch, C.; Cava, M. P.; Brehmer, L.; Ashwell,

G. J. J. Am. Chem. Soc. 1997, 119, 10455-10466. (e) Collier, C. P.; Wong,

E. W.; Belohradsky, M.; Raymo, F. M.; Stoddart, J. F.; Kuekes, P. J.;

Williams, R. S.; Heath, J. R. Science 1999, 285, 391-394. (f) Heath, J.

R.; Kuekes, P. J.; Snider, G. S.; Williams, R. S. Science 1998, 280, 1716-

1721.

As in our previous studies, Al deposition was performed at

room temperature using thermal sources, and analysis was

performed using a combination of in situ surface probes: time-

of-flight secondary-ion mass spectrometry (ToF-SIMS), X-ray

photoelectron spectroscopy (XPS), and infrared reflection

spectroscopy (IRS). In the case of the HO-terminated SAM,

our results indicate that Al chemically interacts with the -OH

groups yielding products that include an H-Al-O-C structure.

Reaction appears to cease after approximately one Al atom is

deposited per molecule. The next about four Al atoms per

molecule form an electron-deficient (premetallic) overlayer with

subsequent deposition forming metallic overlayers. Penetration

of the Al atoms to the S/Au interface is not observed. In contrast,

when Al is deposited onto a H₃CO-terminated SAM, no

chemical reaction with the terminal group is observed; only a

weak complexation occurs. Instead, initial deposition of an

average of three atoms per molecule results in penetration

through the monolayer to the S/Au interface, reminiscent of the

behavior of the H₃C-terminated SAM reported previously.²⁶

Further deposition results in formation of an Al overlayer

(18) Duncan, M. A. Int. J. Mass Spectrom. 2000, 200, 545-569.

(19) Niedner-Schatteburg, G.; Bondybey, V. E. Chem. ReV. 2000, 100, 4059-

4086.

(20) Lisy, J. M. Int. ReV. Phys. Chem. 1997, 16, 267-289.

(21) For example, see: (a) Beierlein, T. A.; Brutting, W.; Riel, H.; Haskal, E.

I.; Muller, P.; Rieb, W. Synth. Met. 2000, 111-112, 295-297. (b)

Campbell, I. H.; Smith, D. L. Appl. Phys. Lett. 1999, 74, 561-563. (c)

Polzonetti, G.; Russo, M. V.; Infante, G.; Furlani, A. J. Electron Spectrosc.

Relat. Phenom. 1997, 85, 73-80. (d) Faraggi, E. Z.; Davidov, D.; Cohen,

G.; Noach, S.; Golosovsky, M.; Avny, Y.; Neumann, R.; Lewis, A. Synth.

Met. 1997, 85, 1187-1190. (e) Vuillaume, D.; Boulas, C.; Collet, J.;

Davidovotis, J. V.; Rondelez, F. Appl. Phys. Lett. 1996, 69, 1646-1648.

(f) Gupta, R.; Misra, S. C. K.; Malhotra, B. D.; Beladakere, N. N.; Chandra,

S. Appl. Phys. Lett. 1991, 58, 51-52.

(22) Jursic, B. S. Chem. Phys. 1998, 237, 51-58.

(23) Di Palma, T.; Latini, A.; Satta, M.; Varvesi, M.; Giardini, A. Chem. Phys.

Lett. 1998, 284, 184-190.

(24) Sakai, S. J. Phys. Chem. 1993, 97, 8917-8921.

(25) Fisher, G. L.; Hooper, A. E.; Opila, R. L.; Jung, D. R.; Allara, D. L.;

Winograd, N. J. Electron Spectrosc. Relat. Phenom. 1998, 99, 139-148.

(26) Hooper, A. E.; Fisher, G. L.; Konstadinidis, K.; Jung, D. R.; Nguyen, H.;

Opila, R. L.; Collins, R. W.; Winograd, N.; Allara, D. L. J. Am. Chem.

Soc. 1999, 121, 8052-8064.

(27) Fisher, G. L.; Hooper, A. E.; Opila, R. L.; Allara, D. L.; Winograd, N. J.

Phys. Chem. 2000, 104, 3267-3273.

9

J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002 5529

A R T I C L E S

Fisher et al.

spectra were acquired using a total ion dose of less than 10¹¹ions/cm².

Relative peak intensities are reproducible to within (6% from both

sample to sample and scan to scan.

Aluminum was deposited onto the sample at room temperature from

a W wire basket source at a rate of ∼0.15 atoms nm^-2s^-1with the

pressure remaining below 5 × 10^-8Torr. After deposition, the

forechamber pressure was allowed to recover to the base value of 1.5

× 10^-9Torr before sample transfer to the analysis chamber. The

deposited mass/area was monitored using a Maxtek, Inc. TM-400 quartz

crystal microbalance (QCM) controller with a maximum error within

(8%.

2.4. X-ray Photoelectron Spectroscopy. The XPS analyses were

performed on a spectrometer (Scienta ESCA 300) equipped with a

monochromatic Al KR source, as described in detail elsewhere.^31,32

A

pass energy of 75 eV and an energy step of 0.05 eV were used for the

analysis. The resulting full width at half maximum (fwhm) for Au 4f_7/2

is 0.52 eV. A binding energy of 84.00 eV for Au 4f_7/2was used as a

reference for all spectra.

Figure 1. Cartoon illustrations of the important features of the interaction

regimes of deposited Al on the HO SAM (top) and the OCH₃SAM (bottom).

The structures depicted are based on common interpretations drawn from

the combined ToF-SIMS, XPS, and IR data. The metal overlayer in the

diagram of the OCH₃SAM is sketched arbitrarily as a smooth film for

purposes of presentation. The detailed morphology remains to be character-

ized.

Following analysis of the bare monolayer, the sample was transferred

under vacuum to the deposition chamber, which is isolated from the

analysis chamber by a gate valve. The pressure in the preparation

chamber remained below 5 × 10^-8Torr during the deposition.

Incremental amounts of Al were deposited at a highly controlled,

constant rate, typically near ∼0.1 atoms nm^-2s^-1, by evaporation from

a graphite crucible. The deposition rate was checked periodically by

removing a reference sample and analyzing with Rutherford backscat-

tering spectroscopy (RBS). After deposition, the Al/SAM specimen

was transferred directly under vacuum to the analysis chamber where

the pressure was maintained below 5 × 10^-9Torr.

growing from the vacuum/SAM interface. A pictorial summary

of the results is shown in Figure 1.

2. Experimental and Methods

2.1. Materials and General Procedures. Ethanol (Pharmco),

hexadecanolide (Sigma Aldrich), acetic acid (¹⁸O₂), and methanol (¹⁸O)

(Isotec) and all other reagents were used as received, except for THF

(Aldrich), which was dried by distillation from Na/benzophenone, and

CH₃OH, which was distilled from Mg turnings. All glass apparatus

was flame dried and purged with dry N₂. Reactions were performed in

a static N₂atmosphere. Flash column chromatography was performed

using 230-400 mesh silica gel, and melting points were taken using a

2.5. Infrared Spectroscopy. Analyses were performed on a Fourier

transform instrument (Mattson Research Series 1000) fitted with custom

in-house optics configured external to the instrument and designed for

grazing incidence reflection of samples under vacuum. A liquid nitrogen

cooled MCT detector was used with an effective low-frequency cutoff

of ∼750 cm^-1. The infrared beam was allowed to access the vacuum

system and reflect from the sample through a pair of differentially

pumped KBr windows. After analysis of the bare monolayer, a shield

was moved to unblock the path between the sample and the Al source.

The Al metal was evaporated from a W-wire basket at a rate of ∼0.15

atoms nm^-2s^-1as measured by a QCM. The pressure remained below

1 × 10^-7Torr during the deposition.

1

Thomas Hoover melting point apparatus. All H NMR spectra were

obtained on a Bruker 200 MHz spectrometer. The aluminum metal for

all depositions (Goodfellow and R. D. Mathis) was of g99.999% purity.

2.2. SAM Preparation. The preparation and characterization of the

types of SAMs used in this study have been described in detail

previously.^28,29The deposition metals (Au and Cr; Goodfellow and R.

D. Mathis, respectively; purities g99.99%) were thermally deposited

sequentially (Cr, ∼5 nm; Au, ∼200 nm) onto clean Si(001) native oxide

covered wafers. Self-assembly of well-organized monolayers was

achieved by immersing the Au substrates into millimolar solutions of

the relevant hexadecanethiol molecules in absolute ethanol for ∼4 days

at ambient temperature. The films were characterized with single

wavelength ellipsometry, infrared spectroscopy, and contact angle

measurements to ensure dense packing and clean surfaces. In addition,

all SAMs were characterized by the initial ToF-SIMS, XPS, and IRS

measurements prior to metal deposition.

2.6. Quantum Theory Calculations. Density functional theory

(DFT) calculations were performed to give estimates of the interaction

energies between Al atoms and the various constituent groups of the

molecules in the SAMs and to show associated changes in the molecular

vibrational spectra. The calculations were done using the algorithms

available in the Gaussain 98 program package.³³The small sytems (less

than five heavy atoms) were geometry optimized using the B3PW91

functional with 6-31G(d,p), 6-31+G(d,s), 6-311+G(d,p), and 6-311+G-

(2d,fp) basis sets.^34,35For the larger systems, the B3PW91/6-31G(d,p)

level of theory was employed. In all cases, frequency calculations were

2.3. Time-of-Flight Secondary Ion Mass Spectrometry. The ToF-

SIMS analyses were performed on a custom designed instrument, as

described previously.³⁰Briefly, the instrument consists of a loadlock,

a preparation chamber, and the primary analysis chamber, each

separated by a gate valve. The primary Ga⁺ions were accelerated to

25 and 15 keV for the HO- and H₃CO-terminated SAM studies,

respectively. The ions were contained in a 100 nm diameter beam that

was rastered over a 1600 × 1600 µm²area during data acquisition. All

(31) Beamson, G.; Briggs, D.; Davies, S. F.; Fletcher, I. W.; Cark, D. T.; Howard,

J.; Gelius, U.; Wannberg, B.; Balzer, P. Surf. Interface Anal. 1990, 15,

541-549.

(32) Gelius, U.; Wannberg, B.; Baltzer, P.; Fellnerfeldegg, H.; Carlsson, G.;

Johansson, C. G.; Larsson, J.; Munger, P.; Vegerfors, G. J. Electron

Spectrosc. Relat. Phenom. 1990, 52, 747-785.

(33) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.

A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann,

R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin,

K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,

R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;

Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.;

Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz,

J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;

Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng,

C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.;

Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon,

M.; Replogle, E. S.; Pople, J. A. Gaussian 98, revision A.9; Gaussian,

Inc.: Pittsburgh, PA, 1998.

(28) Nuzzo, R. G.; Dubois, L. H.; Allara, D. L. J. Am. Chem. Soc. 1990, 112,

558-569.

(29) Laibinis, P. E.; Bain, C. D.; Nuzzo, R. G.; Whitesides, G. M. J. Phys.

Chem. 1995, 99, 7663-7676.

(30) Braun, R. M.; Blenkinsopp, P.; Mullock, S. J.; Corlett, C.; Willey, K. F.;

Vickerman, J. C.; Winograd, N. Rapid Commun. Mass Spectrom. 1998,

12, 1246-1252.

9

5530 J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002

Vapor-Deposited Aluminum Atoms

A R T I C L E S

Figure 3. High-resolution SIMS spectral overlays from the OH SAM. The

spectra show the evolution of metal-organic fragment ions with Al

deposition. (A) Positive ions, nominal mass 70 amu. (B) Negative ions,

nominal mass 127 amu. The spectra are normalized to the initial peak

Figure 2. (A) Negative SIMS spectra showing the 580-820 m/z region of

the OH SAM prior to and following deposition of Al. (B) Integrated peak

+

intensities of C₅H₁₀and C₉H₁₉^-, respectively. (C) Integrated peak areas

-

area of Au₂A^-and AuA₂plotted versus θ_Al.

+

of Al⁺, Al₂⁺, and Al₃plotted versus θ_Al.

performed to ensure the geometry was a true minimum. As a cross

check, we note that our levels of theory give results that compare closely

with the reported results of MP4(SDTQ), MP2, QCISD, and CCSD-

(T) calculations of the energetics of Al-H₂O complexes.^24,36

At least 20 calculations were done for each level of theory to test

the reproducibility of the energies, and several starting geometries for

all the different levels of theory were used to ensure geometry

convergence. To put the results on a more thermodynamic basis, the

energies are reported as enthalpies of the final structures relative to

the isolated reactants and contain zero-point energy corrections and

thermal energy corrections for standard conditions of 1 bar and 298 K.

The thermal corrections to the zero-point energies are in the range of

∼5 kJ/mol.

3.2. Hydroxyl-Terminated Monolayer. 3.2.1. ToF-SIMS.

Negative ions in the mass range from 580 to 820 Da are shown

in Figure 2A. With the deposition of Al, there are significant

changes in the spectra indicating that a reaction occurs between

-

the Al and the OH group. The areas of the Au₂A^-and AuA₂

peaks, involving intact adsorbate molecules, are plotted versus

θ_Alin Figure 2B. The disappearance of these ions with

increasing θ_Alis a direct indication of the conversion of the

original adsorbates to some different chemical structure by the

-

Al atoms. The decrease in the Au₂A^-and AuA₂intensities

also suggests that on average approximately one Al atom per

molecule is required.

3. Results

Evidence that the changes in the adsorbate structures involve

reaction of the OH group with Al to form an Al-O bond can

be deduced from the observation of the formation of AlO-

containing ions. Two of these ions are shown in Figure 3A,B,

where the intensities have been normalized to the initial peak

intensities of C₅H₁₀⁺and C₉H₁₉^-to make obvious the increasing

intensity of the metal-organic fragments with respect to the

simple hydrocarbon fragments as the deposition progresses. The

appearance of the [(CH₂)₆OAl]^-ion fragment (Figure 3B)

indicates that the Al + -OH reaction forms an Al-OR moiety,

where R ) alkyl. We note that the absolute peak intensities of

both the metal-organic and the hydrocarbon fragments reach a

maximum intensity at θ_Al≈ 2-3 atoms/molecule and then begin

to decrease in intensity with increasing deposition (data not

shown). This observation is consistent with a uniformly

distributed aluminum overlayer. The observation of Al₂O⁺

(Figure 3B), again indicative of an Al-OR bond in the reacted

3.1. Definition of Deposited Al Coverage. The deposition

of Al onto the samples was followed directly as the mass per

unit area by either QCM or the integrated flux from a controlled

rate source (XPS case). For convenience in interpretation of

the data, the deposited amounts were converted to coverage in

atoms of Al per SAM molecule, designated θ_Al, on the basis

that there are 4.6 molecules/nm²in a well-formed alkanethiolate/

Au{111} SAM.³Thus for θ_Al) 1.0, there would be one Al

atom deposited on average per SAM molecule.

(34) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652. (b) Perdew, J. P.

In Electronic Structure of Solids; Ziesche, P., Eschrig, H., Eds.; Akademie

Verlag: Berlin, 1991; pp 11-20. (c) Perdew, J. P.; Chevary, J. A.; Vosko,

S. H.; Jackson, K. A.; Pederson, M. R.; Singh, D. J.; Fiolhais, C. Phys.

ReV. B 1992, 46, 6671-6687.

(35) (a) Wachters, A. J. H. J. Chem. Phys. 1970, 52, 1033-1036. (b) Hay, P.

J. J. Chem. Phys. 1977, 66, 4377-4384. (c) Raghavachari, K.; Trucks, G.

W. J. Chem. Phys. 1989, 91, 1062-1065.

(36) Fangstrom, T.; Lunell, S.; Kasai, P. H.; Eriksson, L. A. J. Phys. Chem. A

1998, 102, 1005-1017.

9

J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002 5531

A R T I C L E S

Fisher et al.

adsorbate species, is interesting since fragments of this type have

not been observed in related systems studied to date.^25-27

To establish unambiguously that the appearance of the Al-O

species arises from Al + OH species and not due to reactions

with adventitious sources of oxygen, such as H₂O or O₂, further

analyses were carried out using H¹⁸O-terminated SAMS.³⁷These

data confirm that the deposited Al chemisorbs to the SAM;

signals such as Al₂¹⁸O⁺are observed.

Other metal-organic ions observed in the mass spectra after

Al deposition are AlO^-, AlOC⁺, AlO(CH₂)_x⁽, and AlO(CH)_x-

(CH₂)_y⁽. Of particular interest is the appearance of [Au₂S(CH₂)₁₆]^-

and [Au₂S(CH₂)₁₆OAl]^-ions at nominal masses of 650 and 693,

respectively. These data again indicate that the Al + OH group

reaction results in the formation of Al-OR bonds. We note

that the signal at m/z ) 650, a non-O-containing fragment, is

quite strong in contrast to the O-containing fragment signals at

m/z ) 693 and 70 (Al₂O⁺peak in Figure 3B). These data might

suggest that the C-O adsorbate bond is weakened following

attachment of the Al.

Figure 4. The core-level XPS spectra of the OH SAM prior to and

following deposition of Al. A, B, and C represent the C 1s, O 1s, and Al

2p spectra, respectively.

Evidence that the Al atoms are localized at the top (vacuum

surface selectivity. The binding energies (BEs) were set relative

to the Au 4f_7/2energy at 84.0 eV.

interface) of the SAM, in contrast to penetrating into the

-

molecular matrix, is shown directly by the absence of Au_XAl_YS_Z

,

The C 1s spectra of the bare monolayer show peaks at 285.2

and 287.0 eV corresponding to -CH₂- and -CH₂OH, respec-

tively (Figure 4A). A single O 1s peak at 533.4 eV is assigned

to the -CH₂OH functional group oxygen (Figure 4B). The BEs

of these assignments are close to those reported for poly(vinyl

alcohol) films.³⁹

AlS_X^-, and AlS(CH₂)_X⁺species in the spectra. We have shown

previously for ω-substituted hexadecanethiolate SAMs/Au that

such peaks involving the combinations of Al with S and Au

arise only when the Al penetrates to the S/Au interface, or at

least to a location within ∼0.5-1 nm of the interface.^25-27

The lack of penetration as well as the direct formation of a

Al-OR species are confirmed by the trends in the intensities

of Al⁺, Al₂⁺, and Al₃⁺peaks with Al coverage, shown in Figure

3C. We have shown previously^25-27that these signal intensities

differ significantly between cases where deposited Al atoms

penetrate to the SAM/Au interface (H₃C-terminated) and where

they chemisorb at the vacuum/SAM interface (H₃CO₂C- or

HO₂C-terminated). For example, when Al is deposited onto the

The spectra are noticeably altered upon deposition of Al. The

-CH₂OH C 1s peak in Figure 4A is barely observed as a

broadening of the high BE side of the -CH₂- peak at θ_Al

)

0.7 Al atoms/molecule and has vanished by θ_Al< 1.3. With

increasing deposition, the main -CH₂- C 1s peak shifts slightly

to higher BE (285.7 eV), accompanied by a decrease in signal,

then shifts partially back toward the original value. The initial

shift can be associated with a lowering of the electron density

on the -CH₂-O carbon,⁴⁰while intensity attenuation is

expected from inelastic photoelectron electron scattering in the

Al overlayer. In addition, at low coverages, contributing effects

due to the presence of isolated Al atoms or clusters may arise.⁴¹

The lack of a peak in the region around 282 eV where metal

carbide species have been reported to appear^9,42,43is taken as

evidence for the absence of formation of Al-C bonds, consistent

with our earlier results with other SAMs.^25-27

Examination of Figure 4B shows that the O 1s peak, initially

at 533.4 eV, shifts slightly lower to 532.6 eV and broadens

significantly early in the Al deposition. The BE shift indicates

an increased electron density on the hydroxyl group O atom,

while the broadening suggests the presence of two oxygen

species that are close in energy.⁴⁰When θ_Al> 1.3, the O 1s

peak narrows and shifts to 532.8 eV, a value still lower than

H₃CO₂C-terminated SAM, only Al⁺is observed below θ_Al

1, an indication that the overlayer consists primarily of isolated

≈

Al atoms bonded to organic functional groups. Appreciable

+

levels of Al₂and Al₃are observed only upon further Al

deposition (note the expanded scales in the plots). The appear-

ance of these peaks indicates that reactive sites at the organic

monolayer terminus have been depleted, allowing the Al atoms

to initiate clustering.³⁸The Al_X⁺signals in Figure 3C are similar

to those reported for Al deposited onto a H₃CO₂C-terminated

monolayer (comparison not shown).^25-27Note that the Al⁺

signal in Figure 3C dominates at θ_Al≈ 1, while appreciable

+

levels of Al₂and Al₃appear only at higher coverages.

+

Furthermore, the appearance of Al₂and Al₃in the mass

spectrum does not occur until after θ_Al> 1. We conclude from

these data that approximately one Al atom is bound per hydroxyl

functional group. Indeed, the stoichiometry of Al to hydroxyl

groups, as indicated by the Al_Xcluster data, is in agreement

with that of Figure 2B (Au_XA_Yvs θ_Al).

+

(39) Beamson, G.; Briggs, D. High-Resolution XPS of Organic Polymers: The

Scienta ESCA Database; John Wiley and Sons: New York, 1992; p 94.

(40) Chakraborty, A K.; Davis, H. T.; Tirrell, M. J. Polym. Sci., Part A: Polym.

Chem. 1990, 28, 3185-3219.

(41) In the low Al atom coverage region, the behavior of the spectra may be

influenced by final state effects arising from the presence of isolated Al

atoms and small clusters. While these effects constrain the comparisons of

the BE values of new peaks with standard assignments to be somewhat

qualitative, the internal trends in the observed spectra should remain quite

valid.

(42) Akhter, S.; Zhou, X. L.; White, J. M. Appl. Surf. Sci. 1989, 37, 201-216.

(43) Zhang, P. Study of the Chemistry and Morphology at the Interface between

Metals and Polymers by Spectroscopic Techniques. Ph.D. Thesis, Penn

State University, 1993.

-

3.2.2. XPS. The C 1s, O 1s, and Al 2p spectra of the HO

SAM, before and after deposition of Al, are shown in Figure 4.

A 10° takeoff angle (near grazing) was used to enhance the

(37) A detailed discussion of the isotope labeling experiments is given in the

Supporting Information.

(38) Sigmund, P., Ed. Fundamental Processes in Sputtering of Atoms and

Molecules (SPUT92); Bianco Lunos Bogtrykkeri: Denmark, 1992; pp 223-

254.

9

5532 J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002

Vapor-Deposited Aluminum Atoms

A R T I C L E S

Figure 5. Low- and high-frequency region IRS spectra of the OH SAM

prior to and following deposition of Al. Dashed lines are meant to guide

the eye to follow the evolution of features attributed to the formation of

Al-O and Al-H bonds.

Figure 6. Overlay of the IRS spectra of the bare OH SAM and the same

region of the spectrum deposition of Al to a coverage of θ_Al) 2.3.

assigned as a stretching mode that is unique to the hydroxyl-

terminated monolayer and which is proposed to be associated

with the HO- group mediated conformation of the terminal

-CH₂- unit on the alkyl chain.²⁸

The features related to the OH group are significantly

perturbed upon deposition of Al. In particular, the 1060 cm^-1

C-O stretching mode peak disappears by θ_Alj 0.7, while a

new sharp feature appears at 1095 cm^-1. The changes can be

seen in more detail in Figure 6, which shows an overlay of the

initial spectrum and one after θ_Al) 2.3.

the original one for -OH. This would be consistent with

formation of an Al-O bond. Taken together, the C 1s and the

O 1s data suggest that O atoms of the OH functional groups

react with deposited Al to form an Al-O-C complex with an

approximate average stoichiometry of 1:1.^40,44

The Al 2p spectra in Figure 4C show a peak centered at 75.2

eV when θ_Ale 0.7. The location of this peak, ∼1 eV and ∼1.5

to 2.0 eV higher than expected for Al₂S₃and for Al₂O₃and

Al(OH)_X, respectively, suggests the formation of a new Al

species. The most likely possibility is an organo-aluminum

Because the 1095 cm^-1feature arises in the general C-O

stretching mode region, it is likely that the C-O bond has not

been broken but only perturbed, for example, by Al insertion

into the O-H bond to form a C-O-Al-H species. Aluminum

alkoxides generally have C-O stretching modes in the region

of 1000-1100 cm^-1; for example, Al(OC₂H₅)₃exhibits a strong

species formed by reaction with the OH group. When θ_Al

g

5.3, another peak appears at 72.9 eV, corresponding to metallic

Al. An intermediate phase with a BE of ∼73.5 eV exists

between chemisorption of initially deposited Al (θ_Alj 0.7) and

formation of a metallic overlayer (θ_AlJ 5.3). This phase is

apparently nonmetallic as evidenced by the fact that the metallic

Al peak is located 0.6 eV lower in binding energy.^25-27,42,44

The overall analysis of the data⁴¹indicates that as the

deposition progresses Al atoms first bind with the oxygen of

the OH groups to form an ∼1:1 Al-O-C organometallic

species. Following this, about four more Al atoms deposit to

give a phase with nonmetallic character followed finally at

higher coverages by the formation of a metallic Al overlayer.

C-O stretching mode peak at 1059 cm^{-1 45,46}

.

As the deposition progresses, the lack of changes in this new

band indicates that no further reaction with the -OH group

occurs. In contrast to the C-O stretching mode behavior, the

CH₂bending mode is relatively unaffected by Al deposition.

Note the very strong, broad (fwhm ≈ 200 cm^-1) feature

centered at ∼850 cm^-1that appears when θ_AlJ 1.3 and

intensifies as coverage increases. This band position is typical

of Al-O stretching modes, for example, as seen in the LO mode

47

3.2.3. IRS. The IR spectra of the HO-terminated monolayer,

before and after Al deposition, are shown in Figure 5. The peak

assignments of the bare monolayer have been reported previ-

ously,²⁸and the important features are summarized here for ease

of Al₂O₃and is tentatively assigned to an Al-O species

resulting from reaction of Al with the OH group.⁴⁸

Appearing in concert with the ∼850 cm^-1peak is a broad

(fwhm ) 125 cm^-1) peak at ∼1865 cm^-1. The possibility of

an overtone of the fundamental Al-O absorption is excluded

since that should appear at ∼1700 cm^-1, ∼150 cm^-1lower in

frequency than the observed band. The feature, however, does

fall in the range for an Al-H stretching mode. Typical values

for various organoaluminum hydrides are observed between

1675 and 1925 cm^-1with compounds of the general type HAl-

(R)(OR′) exhibiting peaks in the 1800-1850 cm^-1region.^45,46

of comparison: 1060 cm^-1, C-O stretch (ν_C-O); 1469 cm^-1

,

-CH₂- scissor deformation (γ_CH2); 2850 and 2919 cm^-1, C-H

symmetric (d⁺) and antisymmetric (d^-) stretches of the -CH₂-

chain, respectively. These data indicate that the starting mono-

layer is well organized with the alkyl chains primarily in the

all-trans conformation.²⁸Other small features in the low-

frequency region between the 1060 and 1469 peaks were not

interpreted since the signal level was too close to the noise limit

for useful analysis. The small absorption at 2878 cm^-1is

(45) Maslowski, E., Jr. Vibrational Spectra of Organometallic Compounds;

Wiley: New York, 1977; pp 107-134.

(46) Compton, T. R. ComprehensiVe Organometallic Analysis; Plenum: New

York, 1987; pp 283-308.

(44) Stoyanov, P.; Akhter, S.; White, J. M. Surf. Interface Anal. 1990, 15, 509-

515.

(47) Mertens, F. P. Surf. Sci. 1978, 71, 161-173.

9

J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002 5533

A R T I C L E S

Fisher et al.

Figure 7. Integrated SIMS peak areas of Au₂A^-and AuA₂^-plotted versus

θ_Alfor the OCH₃SAM.

We note though that these bands are typically much narrower

than the observed 1865 cm^-1feature. This indicates significant

inhomogeneous broadening in the stretching mode of the surface

compounds relative to pure organoaluminum compounds. Steric

crowding of the relatively large organo-aluminum species

formed at the monolayer termini would be expected to lead to

a distribution of molecular configurations that would give rise

to inhomogeneous broadening. Such effects also seem likely

as interpreted from the C-H stretching mode changes noted

below.

The high-frequency region of the spectrum indicates that the

symmetric and antisymmetric C-H stretches shift to higher

frequency and broaden by θ_Al≈ 1.3 Al atoms/molecule with

the d^-mode shifted upward by 6 cm^-1at θ_Al≈ 2.3. These

results indicate that the conformational ordering of the alkyl

chains decreases when Al is deposited. This effect is expected

on the basis that steric crowding of the organo-aluminum species

would disrupt chain packing via creation of gauche disorder

which would propagate down the chains away from their

termini.^25-27

Figure 8. High-resolution SIMS spectral overlays from the OCH₃SAM.

The spectra show the progression of metal-organic fragments with Al

deposition. A, B, and C represent, respectively: positive ions, nominal mass

58 Da; positive ions, nominal mass 61 Da; and negative ions, nominal mass

453 Da. The intensities in plots A and B are normalized to the initial peak

intensity of C₄H₁₀⁺and C₂H₅S⁺, respectively, whereas the spectra in C are

normalized to the peak intensities of Au^-.

the first increment of Al, the increase of the Au₂A^-and AuA₂

-

peak intensities (Figure 7) is ascribed to electron transfer from

electropositive Al atoms to the more electronegative Au atoms

and clusters leaving the surface. As the deposition progresses,

all peak intensities become increasingly attenuated, consistent

with a growing Al overlayer that can block substrate ion ejection.

In the positive ion mass spectrum, there is evidence that Al

interacts with the terminal -OCH₃group. In Figure 8A, the

3.3. Methoxy-Terminated Monolayer. 3.3.1. ToF-SIMS.

After deposition of Al, an important diagnostic feature in the

ToF-SIMS spectra, shown in Figure 7, is the relatively constant

intensity of the Au₂A^-and AuA₂^-peaks, which involve intact

adsorbate molecules. Note in Figure 7 how the areas of these

peaks barely drop below their initial values in the bare SAM at

continued Al deposition. These data indicate that deposition of

Al leaves the adsorbate molecule chemically intact. Thus there

is no reaction between the Al atoms and the OCH₃groups.

Consistent with this interpretation, the hydrocarbon fragment

peak intensities remain relatively unchanged during the early

stages of the deposition (data not shown). Upon deposition of

+

intensity of the AlOCH₃(m/z ) 58) peak is shown for

increasing values of θ_Al. The spectra are normalized to the initial

+

peak intensities of C₄H₁₀to make obvious the changing

intensities of the peaks with respect to the hydrocarbon and

(

substrate fragments as the deposition progresses. No Al_XO_Y

ions were observed. On the basis of our previous work with

-CO₂H²⁷and -CO₂CH₃,^25-27the appearance of AlOCH₃⁺ions

(48) There has been significant controversy over exact assignments of the

positions of Al-OC stretching modes in aluminum alkoxides [ref 46] with

frequencies ranging from ∼500 to 1000 cm^-1. The assignments have been

complicated in some cases by the presence of Al-O-Al species with

associated Al-O stretching modes. At θ ≈ 1, the band is taken as evidence

for formation of an Al-O bond involving the OH group. While the Al-O

stretching cross section is known to be generally intense, for example, in

the case of the Al₂O₃phonon [ref 47], we note that the intensity of our

observed mode at higher coverages is considerably stronger than would be

expected from comparison to a similarly assigned Al-O stretching mode

features in previous studies of Al deposition on HO₂C- and H₃CO₂C-

terminated alkanethiolate SAMs on Au [refs 26, 27]. The formation of Al₂O₃

by reaction of Al with adventitious O₂is not the cause based on ¹⁸OH

ToF-SIMS data (see Supporting Information). Furthermore, under the

identical experimental conditions, Al deposition on other SAMs does not

give rise to the intense 850 cm^-1feature [refs 25-27]. One possibility for

these intensity variations would be variations of the oscillator orientations

in the different cases in which the intensity would increase with increasing

vertical orientation of the transition dipole.

(

but not Al_XO_Yions indicates that the deposited Al has not

undergone an insertion interaction with -OCH₃to form Al-O

bonds, and suggests that a weak organo-aluminum complex has

formed. This is also consistent with the observed behavior of

-

the AuA₂and Au₂A^-ions, which indicates that there is no

reaction between the terminal group and Al.

Furthermore, there is also evidence for the simultaneous

penetration of the Al atoms to the S/Au interface. In Figure 8B

and C, the intensities of the AlSH₂⁺(m/z ) 61) and the Au₂AlS^-

(m/z ) 453) cluster peaks are shown for increasing Al

deposition. The spectra are normalized, respectively, to the initial

peak intensities of C₂H₅S⁺and Au^-to make obvious the

9

5534 J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002

Vapor-Deposited Aluminum Atoms

A R T I C L E S

Figure 9. Integrated SIMS ion peak areas of plotted Al_n⁺, n ) 1-3 versus

θ_Alfor the OCH₃SAM.

changing intensities of the peaks with respect to the hydrocarbon

and substrate fragments as the deposition progresses. From our

earlier work with the H₃C-terminated SAM,^25,26the observation

of ions containing a combination of Al and S and/or Au indicates

that the deposited Al has penetrated through the CH₃O-

terminated SAM to the S/Au interfacial region, where the

formation of mixed cluster peaks becomes possible (see earlier

discussion in section 3.3.1). The intensity of the AlSH₂⁺peaks

is lower than expected if all the deposited Al penetrated through

the SAM to the Au/S interface when compared to our observa-

tions of Al deposited on an H₃C-terminated SAM. This reduced

intensity is consistent with the deposited Al both penetrating to

the Au/S interface and simultaneously associating with the OCH₃

group. We also note that the increasing relative Au₂AlS^-

intensity (Figure 8C) throughout the deposition regime suggests

that Al penetrates through the SAM to the Au/S interface at all

coverages.

Figure 10. The core-level XPS spectra of the OCH₃SAM prior to and

following deposition of Al. Plots A, B, and C represent the C 1s, O 1s, and

Al 2p spectra, respectively.

that an organo-aluminum complex is formed between the

terminal group and deposited Al at all coverages. At the same

time, we observe the formation of cluster ions between Al and

S and/or Au indicating that the deposited is simultaneously

penetrating to the Au/S interface.

3.3.2. XPS. The C 1s, O 1s, and Al 2p core-level shifts for

the H₃CO-terminated SAM are shown in Figure 10. The data

were collected under similar conditions to that of the HO-

terminated SAM, except that a 15° takeoff angle was used

instead of 10°. The C 1s spectra peaks at 284.9 and 286.8 eV

are ascribed to the -(CH₂)- alkyl chain and the -CH₂OCH₃

C atoms, respectively (Figure 10A). The single O 1s peak at

533.1 eV is assigned to -CH₂OCH₃(Figure 10B). These

assignments are in good agreement with those reported for poly-

(vinyl methyl ether).⁴⁹

Information on the state of the penetrated Al atoms can be

+

seen by examining the Al_Xion signals, shown in Figure 9.

These data show that with the first increment of deposited Al,

the monomer, dimer, and trimer intensities increase proportion-

ately. We have demonstrated earlier that these signals differ

between the case where deposited Al chemisorbs at the

monolayer terminus and where it penetrates through the

monolayer to the S/Au interface.^25-27When Al is first deposited

onto a H₃C-terminated SAM, which allows penetration, the Al⁺

The absence of any features near 282 eV indicates that no

aluminum carbide species form.^9,42,43The vanishing of the

-CH₂OCH₃C 1s peak, at 286.8 eV (Figure 10A), by θ_Al

<

1.3 Al atoms/molecule indicates an ∼1:1 perturbing interaction

between Al and the SAM that lowers these BEs, merging them

into the main -CH₂- peak. In turn, the -CH₂- C 1s peak

broadens and shifts to slightly higher binding energy (285.3 eV)

as the deposition progresses.⁴¹

During the initial stages of Al deposition, the 533.1 eV O 1s

peak broadens and loses intensity, while a second peak at 532.5

eV appears and increases in intensity. By θ_Al≈ 5.3 atoms/

molecule, the two peaks combine into one broad peak centered

at 532.8 eV. Together with the C 1s data, this indicates that the

Al atoms interact with the O atoms of the OCH₃. Further

analysis is complicated by the formation of what appears to be

an aluminum oxide species, presumably caused by reaction with

O₂and/or H₂O background gases.⁵⁰

+

and Al₂increase steadily, whereas there is a slight delay in

the Al₃⁺ion. This trend is similar to that for Al deposited onto

a bare, clean Au substrate, consistent with the conclusion that

the H₃C-terminated SAM allows Al to penetrate completely

through the monolayer. In contrast with a reactive terminal

group, such as CO₂CH₃²⁶or OH (section 3.2.1), when the first

increments of Al are deposited on the monolayer, where it

chemisorbs, there is an increase in the Al⁺intensity, while there

is a slight delay in the Al₂⁺and Al₃⁺that rise in tandem (e.g.,

+

see Figure 3C). The early growth of Al₂intensity supports

the conclusion that Al penetrates through the OCH₃SAM to

the Au/S interface.

The Al 2p spectra show the appearance at the higher

coverages of two peaks centered at 75.2 and ∼74 eV. The latter

is close to one observed in our previous work on the H₃C-

terminated SAM²⁶and would be consistent with an Al-S

species. This assignment in the present case is supported by

Thus, the OCH₃SAM appears to exhibit intermediate

reactivity to Al when compared to the CH₃SAM, where Al

penetrates through the layer, and the CO₂CH₃, COOH, and OH

SAMs, where Al chemisorbs at the SAM terminus group. The

absence of Al_XO_Y⁽ions, in contrast to the case of the OH SAM,

establishes that Al-O bonds do not form, even at high Al

coverage where deposition is at the vacuum/SAM interface.

However, the formation of AlOCH₃⁺ions (Figure 8A) indicates

(49) Beamson, G.; Briggs, D. High-Resolution XPS of Organic Polymers: The

Scienta ESCA Database; John Wiley and Sons: New York, 1992; p 90.

(50) By the time these high coverages are delivered in the XPS experiments,

the samples have been exposed for much longer times than in the ToF-

SIMS and IRS experiments (∼10 h vs 1 h).

9

J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002 5535

A R T I C L E S

Fisher et al.

Figure 12. Low- and high-frequency region IRS spectra of the OCH₃SAM

as a function of Al coverage. The inset shows a plot of the C-O stretching

mode integrated intensity as a function of the Al coverage.

(∼17 Å thick). As expected for an ambient surface layer, the C

1s and O 1s curves for the -OCH₃group drop together

identically with increasing angle and exhibit the fastest falloff

of all the data. The curve for the -CH₂- signal drops off much

more slowly, in accordance with the carbon atoms being

distributed in a region spanning the Au surface and the ambient

surface. The data for the Al 2p are divided into the signal for

the metallic (zerovalent) signal and the signal attributed to the

nonmetallic Al-S and/or Al-O species.⁵¹Note that the Al 2p

metallic peak closely follows the dropoff with angle shown by

the surface OCH₃data. This establishes that metallic Al is

deposited primarily at the ambient/SAM surface. In contrast,

the Al-O/Al-S peaks closely follow the alkyl chain (-CH₂-)

layer falloff. This implies that these species are found distributed

both at the surface and at the Au/S interface and is consistent

with the formation of an Al-S species at the Au surface for Al

atoms that penetrate the SAM and some formation of Al-O

species at the ambient surface. We note that these conclusions

are to be regarded as somewhat qualitative due to uncertainties

in the exact assignments of the Al reaction products at low

coverages.⁴¹

Figure 11. Plots of the XPS peak intensities versus the photoelectron

takeoff angle for the OCH₃SAM before and after Al deposition to a

coverage of θ_Al) 5.3. A normalized intensity is plotted as defined by (I/

I_Au4f)_θ/(I/I_Au4f)₁₅, where I and I_Au4f, respectively, are the selected core-level

intensity and the corresponding Au 4f peak intensity for the same spectral

acquisition. The term (I/I_Au4f)_θrepresents the ratio for a spectra recorded at

the takeoff angle of interest, and (I/I_Au4f

)

represents the ratio for spectra

15

taken at 15°. Spectra were taken for O 1s, -OCH₃(O); C 1s, -OCH₃(0);

C 1s, -CH₂- (4); Al 2p, Al (metallic) (b); Al 2p, (Al_XO_Y+ Al_XS_Y) (9);

Al 2p, total Al intensity from all peaks (2). For details see text.

the ToF-SIMS data (see section 3.2.1) that show that some

fraction of the deposited Al atoms penetrates to the S/Au

interface. The peak observed at 75.2 eV is at a higher energy

than that expected for aluminum oxide. In our previous studies,

a peak at this energy was assigned to Al interacting with an O

of -CO₂H and -CO₂CH₃groups.^25,26By analogy, we assign

the peak to be associated with Al coordinated with the OCH₃

group in some fashion.

Both the 74.2 and the 75.2 eV peaks continue to grow in

intensity until θ_Al≈ 2.7. At this point, the 74.2 eV peak starts

to attenuate with further Al deposition, while the first signs of

metallic Al arise at 72.9 eV. Further analysis is complicated,

as for the O1s spectra, due to the formation of what appears to

be background-induced aluminum oxide.⁵⁰

To determine in more detail if Al atoms penetrate to the S/Au

interface as well as interact with the surface OCH₃groups, a

set of variable takeoff angle spectra was obtained at θ_Al) 5.3.

The results are shown in Figure 11, where the selected core-

level peak areas are plotted as the area ratio to the Au 4f peak

at a given takeoff angle normalized to the same ratio at the 15°

grazing angle [(I/I_Au4f)_θ/(I/I_Au4f)₁₅]. The data can be understood

in a qualitative way on the simple basis that at 15° the

photoelectrons sampled have a much higher fraction coming

from atoms at the ambient surface as compared to the fraction

at higher angles where the sampling shifts toward the deeper

regions of the sample. The plot shows three calibration curves

(dashed lines): O 1s and C 1s data for the ambient surface OCH₃

group layer and C 1s for the alkyl chain -CH₂- group layer

3.3.2. IRS. The IR spectra of the H₃CO-terminated mono-

layer, before and after Al deposition, are shown in Figure 12.

The previously reported peak assignments^29,52,53of the bare

monolayer are summarized here for reference. The 1132, 1392,

and 1468 cm^-1peaks are assigned as the C-O-C antisym-

metric stretch (ν_C-O), the -CH₃symmetric deformation (δ_CH),

3

and the -CH₂- scissor deformation (γ_CH), respectively. The

2

-CH₂- d⁺and d^-stretches are assigned at 2851 and 2918

cm^-1, respectively. The peaks at 2811, 2828, and 2981 cm^-1

have been assigned to the various stretching modes of the

terminal CH₃group. The features at 2811 and 2828 cm^-1are

assigned as -CH₃Fermi resonance modes. These data indicate

that the starting monolayer is well organized with the chains

primarily in all-trans conformations.²⁹

The low-frequency region shows that features related to the

OCH₃and the CH₂groups are only slightly perturbed upon

deposition of Al. First, the ν_C-Ostretch mode at 1132 cm^-1

decreases in intensity slightly but remains relatively unchanged

(51) The intensities of these peaks were first corrected for the presence of a

weak Au5p feature at ∼75 eV.

(52) Ong, T. H.; Davies, P. B.; Bain, C. D. Langmuir 1993, 9, 1836-1845.

(53) Allan, A.; McKean, D. C.; Perchard, J.-P.; Josien, M.-L. Spectrochim. Acta

1971, 27A, 1409-1437.

9

5536 J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002

Vapor-Deposited Aluminum Atoms

A R T I C L E S

detailed discussion will be made with reference to the sequence

of major mechanistic paths shown below.

Figure 13. Overlay of the IRS spectra of the OCH₃SAM before (s) and

after (- - -) deposition to a coverage of θ_Al) 2.3. Inset: Plot of the fractional

integrated intensity of the C-O stretching band of the OCH₃SAM as a

function of Al coverage.

In this mechanism, the fate of the deposited Al atoms is

controlled by the competition between local stabilizing inter-

actions (complexation) with a surface group, chemical bond

insertion (C-C, C-H, C-O, and O-H, as appropriate),

penetration to the interior Au/S interface of the SAM, and

nonspecific surface adsorption and film nucleation. For con-

venience, the monolayer is written as SAM-X, where X is OH

or OCH₃in this study.

The sequence of processes is viewed as starting from

adsorption of a deposited Al atom at a random site [Al(s) in

process A] on the SAM surface. While this process should be

considered as generally reversible, the combination of ToF-

SIMS and XPS data in this study does not indicate any

desorption of Al atoms from the SAM surfaces.⁵⁴Surface

diffusion allows access to locations where processes C, I, and

P can occur. The complexation process is written as reversible,

characterized by forward and reverse rate constants, k^x_Cand

k^x_-C. The bond insertion and penetration steps are written as

irreversible with forward rate constants k^x_I-ABand k_P^x, respec-

tively. Penetration is divided into sub paths for adsorption onto

a Au lattice site with minimal S atom interaction and complete

insertion into a Au-S bond. Nonspecific nucleation (process

N) is written purposely to be general and collects together all

deposited Al atoms that are not included in the above categories

at all stages of Al coverages.

until θ_Al≈ 2.7 Al atoms/molecule whereafter the only effect is

increasingly rapid intensity attenuation. A more detailed look

for θ_Al) 2.3 is given in the overlay with the initial bare

monolayer spectrum in Figure 13. The C-O stretching mode

is observed to shift upward by 4 cm^-1with the Al deposition

and appears to sharpen. The change in C-O peak area with θ

is shown in the inset in the figure. The sharpest change is at

θ_Al≈ 0.5, whereas thereafter the effects of Al level off until

about two Al atoms/OCH₃.

Second, the δ_CHand γ_CHfeatures remain essentially un-

3

2

affected upon deposition of Al. These observations indicate that

there is no significant chemical interaction of the OCH₃group

by Al. Furthermore, in contrast to the OH SAM case (see

earlier), there is no observation of an Al-O stretch in the

spectra, expected at ∼855 cm^-1. This establishes that no bond

cleavage takes place at the methoxy group to form metal-oxide

species.

The high-frequency region of the spectrum in Figures 12 and

13 indicates that the d⁺and d^-CH₂C-H stretches, as well as

other bands in the high frequency region, become only slightly

perturbed upon deposition of Al. In fact, the only observed shift

is with the d^-mode (2921 cm^-1) which moves to higher

frequency by only ∼1 cm^-1(see Figure 13). These data establish

that the conformations of the alkyl chains are virtually undis-

turbed by the deposited Al.

The discussion will proceed by considering the two SAMs

in turn, emphasizing the mechanistic aspects of each case.

Finally, the overall energetics and structures of species formed

by the Al-SAM interactions will be examined.

4.1. HO-Terminated MonolayersSurface Trapping via

OH Chemical Reaction. 4.1.1. Overall Process Sequence. The

data indicate that the initial stage of the Al deposition is

dominated by insertion of an adsorbed Al atom into an O-H

4. Discussion

The schematic summary diagram in Figure 1 serves as a

convenient general background picture for the discussion. The

bond (reaction I-OH) to form an O-Al-H species. This implies

OH

I-OH

. k^O_P^H. Neither C-H and C-C bond reactions nor

(54) There are reported cases of desorption of metal atoms’ vapor deposited

onto organic substrates, typically observed for cases of noble metals on

polymers with very low surface energies (see, for example: Zaporojtchenko,

V.; Behnke, K.; Strunskus, T.; Faupel, F. Surf. Interface Anal. 2000, 30,

439-443. Thran, A.; Kiene, M.; Zaporojtchenko, V.; Faupel, F. MRS Bull.

1999, July, 3). In the case of the OH and OCH₃SAMs of the present study,

desorption does not appear to occur as determined by comparing QCM

response with ToF-SIMS and XPS results. Specifically, the ToF-SIMs

and XPS signals that reveal the presence of Al monotonically track the

QCM coverages, which involve crystals precoated with metal films that

give unity sticking coefficients. This is particularly important at low Al

coverages where the highest probability of desorption will occur.

that k

penetration of Al atoms to the Au/S interface is observed.

Following consumption of about one Al atom per group, a

premetallic phase of Al nucleates and grows for about two layers

until growth of a metallic phase initiates.

4.1.2. O-H Bond Insertion Process. Strong evidence for

Al-induced cleavage of the O-H bond is given by the ToF-

SIMS and IR data (Figures 3 and 5). In particular, the

9

J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002 5537

A R T I C L E S

Fisher et al.

Scheme 1

observation of a new feature in the IR spectrum at ∼1850-

1900 cm^-1, assigned to an Al-H stretching mode, supports Al

insertion into the CO-H bond. These data are fully consistent

with the XPS spectra (Figure 4).

The positive shifts (∼6 cm^-1for the d^-mode) and the

broadening of the C-H stretching mode peaks (seen in detail

in Figure 6) indicate that significant conformational disordering

of the alkyl chains occurs in concert with the O-H insertion.

This behavior is consistent with the formation of a reaction

product at the HO- terminal group that gives rise to steric

crowding with increases in the fraction of gauche defects near

the SAM surface.

The data indicate ∼1:1 Al/OH stoichiometry for the O-H

insertion (reaction I-OH). The ToF-SIMS data show the

complete loss of mass peaks for clusters containing intact

adsorbate by θ_Al) 1.0 ((0.15) (Figure 2B). Further, in Figure

3A,B, note the biggest change in the appearance of the Al₂O⁺

and [AlO(CH₂)₆]^-peaks occurs by θ_Al) 1.0. The XPS data

(Figure 4A) show complete loss of the -OCH₃C 1s peak by

θ_Al) 1.3. The IRS data (Figure 5A) show a near cessation of

changes in the C-O stretching peak region by θ_Al≈ 0.7.

Thereafter, a small residual feature remains, shifted slightly to

higher frequencies than the original C-O feature. These data

suggest that ∼15-20% of the OH groups do not undergo Al

insertion but are simply perturbed in some subtle way.⁵⁵

appears very nonlinear with coverage and continues on into the

metallic Al region (θ_Al≈ 5), while little change is seen in the

band shape and peak maximum. This observation suggests that

there is reorientation of the initially formed Al-O bond in a

direction normal to the metal surface, but in the absence of

additional characterization of the nonmetallic aluminum-orga-

nometallic phase(s), the underlying reason for the behavior of

the Al-O feature remains unclear.

4.2. H₃CO-Terminated MonolayersPartitioning between

Overlayer Film Nucleation and SAM Penetration. 4.2.1.

Overall Process Sequence. In complete contrast to the HO-

terminated monolayer system, for which bond insertion at the

terminal group dominates, the data for the H₃OC-terminated

SAM show that the Al atoms exclusively partition between

overlayer film nucleation and penetration to the Au/S interface

(processes N and P) with no chemical bond breaking. The

absence of Al insertion into the OCH₃group is directly

supported by the nearly constant character of the C-O stretching

mode at 1132 cm^-1in the IRS spectra (Figure 12). Other data

are fully consistent with the lack of any chemical reactions with

C-H, C-O, and C-C bonds. The lack of a rapid, irreversible

insertion reaction of the deposited Al atoms at the SAM surface

opens the penetration pathway.

4.1.2. Formation of a Premetallic Al Phase at Intermediate

Coverages. In general, with increasing Al coverage above θ_Al

≈ 1.0, no further reactions of the -OH group appear to occur.

The Al 2p XPS data show (Figure 4) that metallic overlayers

finally form at θ_AlJ 5.3, while for θ_Alj 2.7, or slightly higher,

a positive valence or otherwise nonmetallic phase forms. A

similar nonmetallic aluminum-organic phase was observed prior

to the metallic one in previous studies of the HO₂C- and H₃-

CO₂C-terminated SAMs. In the present case, after cessation of

reaction I-OH, the next about two Al atoms continue to form

nonmetallic products. It is in this coverage regime that the 850

cm^-1peak assigned to the Al-O IR stretching mode begins to

grow in intensity (Figure 5). Note that the intensity growth

The degree of partitioning between nucleation and penetration

is controlled by the strength of stabilization of adsorbed Al

atoms at the surface via localized Al‚‚‚OCH₃complexation

(55) One possible explanation for the appearance of the second Al-OH weak

interaction mode is that each event of reaction 1 produces an organoalu-

minum surface species that occupies more surface area on the SAM than

on the original chain terminus. As the deposition proceeds, the formation

of adjacent product species will cause disruption of the surface packing

and thus eventually could lead to steric screening of the final fraction of

unreacted OH groups with a lowered probability of reaction 1. This

explanation, however, does not seem likely since in the case of Al atoms

deposited on a H₃CO₂C-terminated SAM, the reaction stoichiometry for

reaction of the terminal group was very close to 1:1. A second possibility

for incomplete reaction is that Al reactivity depends on the H-bonding state

of the OH group. It is known that a major fraction of the terminal OH

groups in these types of SAMs is H-bonded to neighbors [Atre, S. V.;

Liedberg, B.; Allara, D. L. Langmuir 1995, 11, 3882-3893]. Thus it is

possible that the small fraction of nonbonded OH groups could be unreactive

towards Al. Approximating the OH groups as being positioned on a (x3

× x3) hexagonal lattice with H-bonding in pairs, then ∼14% of the groups

would be isolated, and thus unreactive with Al. While this fraction is close

to what the data suggest, the weak reactivity of the isolated OH groups

seems unlikely given the highly favorable thermodynamics for formation

of Al-O and Al-H bonds [Chen, J. G.; Basu, P.; Ng, L.; Yates, J. T., Jr.

Surf. Sci. 1988, 194, 397-418. Rogers, J. W., Jr.; White, J. M. J. Vac. Sci.

Technol. 1979, 16, 485-488. Rogers, J. W., Jr.; Hance, R. L.; White, J.

M. Surf. Sci. 1980, 100, 388. Basu, P.; Chen, J. G.; Ng, L.; Colaianni, M.

L.; Yates, J. T., Jr. J. Chem. Phys. 1988, 89, 2406-2411. Waddil, G. D.;

Kesmodel, L. L. Surf. Sci. 1987, 182, L248. Kerkar, M.; Hayden, A. B.;

Woodruff, D. P.; Kadadwala, M.; Jones, R. G. J. Phys.: Condens. Matter

1992, 4, 5043. Stone, F. G. A., West, R., Eds. AdVances in Organometallic

Chemistry; Academic Press: New York, 1996; pp 40-50. Coates, G. E.,

Green, M. L. H., Wade, K., Eds. Organometallic Compounds, 3rd ed.;

Methuen & Co.: London, 1967; Vol. 1, pp 295-343], relative to the weak

energies of intermolecular H-bonds.

(process C-OCH; with associated rate constants, k^O_C^CHand

3

OCH₃

-C

k

). Stabilization reduces access to the penetration channel

by reducing surface mobility and enhances overlayer nucleation

by maintaining a higher surface Al atom population. With

continued Al deposition, the penetration process eventually

closes, while overlayer metallic film growth continues.

The essence of the overall mechanism is summarized in

Scheme 1.

4.2.2. SAM Penetration versus Overlayer Nucleations

Comparison with Previous Results of a H₃C-Terminated

SAM. In both our previously studied case of a H₃C-terminated

SAM²⁶and the present case of the H₃OC-terminated one, the

data show that Al penetrates to the S/Au interface to form a

smooth, uniform interfacial adlayer, leaving the chain confor-

mational ordering virtually unperturbed. There is an important

difference between the two cases, however. For the OCH₃SAM,

-

the ToF-SIMS data show that the mixed Al_XAu_YS_Zcluster

ion peaks, which signal diffusion of Al to the Au/S interface,

continue to increase even up to θ_Al≈ 12.2 (Figure 8B,C). In

contrast, for the CH₃case, they stop at θ_Al≈ 2.7, signaling the

9

5538 J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002

Vapor-Deposited Aluminum Atoms

A R T I C L E S

Table 1. Theoretically Calculated Standard Enthalpy Values for the Interaction of an Isolated Al Atom with an Isolated Model Molecule

molecules and Al interaction products

reaction path

∆H°(298), kJ/mol

experimentally observed

conclusion/comment

Al + CH₃CH₂OH

-C-Al-OH

I_CO

I_OH

C_OH

-248 ((9)

-188 ((6)

-44 ((6)

no

yes

kinetic barrier too high

-O-Al-H

Al‚‚‚O(H)(CH)₂-

Al + CH₃CH₂CH₃

-C-Al-CH₃

I_CC

I_CH

C_CH

-82 ((7)

-44 ((14)

∼0 ((5)

no

kinetic barrier too high

-C-Al-H

Al‚‚‚[CH₃CH₂CH₃]

Al + CH₃OCH₃and CH₃(CH₂)₄OCH₃

-C-Al-OCH₃

I_CO

C_OCH

-263 ((8)

-38 ((5)

no

yes

kinetic barrier too high

-O-Al-CH₃

Al‚‚‚O(CH₃)(CH₂)

3

end of penetration. Past this point, Al overlayer growth at the

H₃C/vacuum interface proceeds.²⁶Other aspects of the ToF-

SIMS and XPS data confirm these differences (see sections 3.3.1

and 3.3.2). For example, the XPS data in Figure 11 show that

at the intermediate coverage of θ_Al) 5.3, the fraction of Al

located at the outer surface is dominantly a metallic phase. We

analyze these differences in terms of the mechanism in Scheme

1.

quantum theory calculations were carried out for the simple

isolated systems of Al + CH₃CH₂OH, CH₃CH₂CH₃, CH₃OCH₃,

and CH₃(CH₂)₄OCH₃. Geometry optimization and energies were

calculated using a variety of geometrical starting points to

facilitate finding the global energy minima of various spatial

configurations. The results are summarized in Table 1, where

the minimum energies are given in terms of ∆H°(298). The

reported values represent the average of the results for the four

basis sets used (see section 2.6), and the errors show the spread

in values.

The calculations show that bond insertion is energetically

favorable for all the molecular bonds with C-O insertion the

most favorable. Because only O-H bond insertion (I-OH) is

observed, it appears that the activation barriers for insertion into

the other bonds are significantly higher than that for the O-H

bond.

Secondary minima at -44 and -38 kJ/mol are found

involving primarily a complexation type of interaction between

Al and the O atom of the -OH and -OCH₃groups, respec-

tively. In the case of the former, the complex can be viewed as

a transient precursor state to a final O-H bond insertion

product.⁵⁸In the case of the latter, the data further show that

association of an Al with the hydrocarbon unit is thermoneutral

thus giving a preference of ∼159 kJ/mol for an Al atom to be

located at O rather than C or H atoms of an alkyl ether molecule.

Reported theoretical and experimental values for similar systems

support the Al‚‚‚O results. For example, calculations [HF/32-

1G, HF/63-1G(d), and MP2/6-31G(d)] for the Al‚‚‚OH₂and Al‚

‚‚O(CH₃)₂complexes²⁴give stabilization energies of -29 to

33 and -38 kJ/mol, respectively, where the energies include

zero-point corrections but not thermal energies. Inclusion of the

latter lowers the values by ∼6 kJ/mol. An experimental

stabilization energy of -38.4((2.5) kJ/mol has been reported

for the Al‚‚‚O(CH₃)₂complex.⁵⁹

First, we set the penetration rate constants for the two SAMs

equal, viz., k^O_P^CH) k_P^CH, on the following basis. It was

proposed²⁶that Al penetration into the CH₃SAM occurs

predominantly via transient defects that arise from thermally

activated hopping of the S atoms of the alkanethiolate chains

away from their lowest energy positions on the Au surface. Such

dynamic fluctuations have been demonstrated theoretically for

alkanethiolates on Au(111) and at room temperature are far more

rapid than the deposition rates and analysis time in our

experiments.^56,57Because both the CH₃and the OCH₃SAMs

have virtually identical SAM/Au interfacial structures, chain

packing densities, chain organization, and intermolecular inter-

actions, it is reasonable to assume that the transient defect

3

fluctuation processes and their associated penetration rate

constants are nearly identical in the two SAMs, viz., k_P^OCH

≈

3

k^C_P^H

.

3

Next we consider the coverage dependence of the penetration

processes. It was previously concluded²⁶that Al penetration into

the CH₃SAM continues until an ∼1:1 Al:Au adlayer is formed,

which for an average Au(111) texture with ∼13 Au atoms/nm²

SUB

Al

would exhibit a substrate saturation coverage θ

≈ 2.8 Al

atoms per molecule. The actual reported saturation coverage of

∼2.7 (∼12.3 Al/nm²) suggests slightly less than a saturation of

1:1 Al/Au adlayer.²⁶Given the near identical SAM structures,

we set θ^SUB(OCH₃) ) θ^SUB(CH₃) ≈ 2.7 Al atoms per molecule

Al

as the closure point for the penetration channel in both SAMs.

On the basis of the mechanism in Scheme 1, the difference

in the overall values of θ_Alrequired to close the penetration

step for different SAMs depends on the competition between

the complexation and penetration channels for Al(s). For

stabilization of the Al atoms, the associated equilibrium constant

While the experimental data indicate formation of a simple

Al‚‚‚O(CH₃)R complex, a measure of the interaction strength

is difficult to ascertain. For example, the ToF-SIMS and IRS

(56) Bhatia, R.; Garrison, B. J. Langmuir 1997, 13, 765-769.

(57) Bhatia, R.; Garrison, B. J. Langmuir 1997, 13, 4038-4043.

(58) Sakai (ref 24) has reported calculations of the transition state barriers of

Al insertion in H₂O and (CH₃)₂O. The transition state energies are ∼33-

46 and ∼88 kJ/mol for H-Al-OH and H₃C-Al-OCH₃insertion,

respectively, relative to the total energy of the isolated reactants. With Al‚

‚‚molecule complexes as precursors, the activation barriers will be raised

by the stabilization energy of the corresponding complexes. Sakai notes

that for the (CH₃)₂O case the probability of crossing the insertion barrier

will be lowered because of orbital overlap effects, making the C-O insertion

process, relative to the O-H insertion, less probable than that predicted

from energetics alone.

for the reaction, written in the direction of complexation of an

x

Al(s), is K^x_C) K^x_C/K_-C. Given that k_P^OCH≈ k_P^CH(see above),

3

then to have a comparatively diminished penetration channel

in the OCH₃case, k^OCH> k_C^CH; that is, the OCH₃group acts

3

C

as a better Al atom trap or penetration gate than does CH₃.

4.3. Energetics and Structures of the Al + SAM Molecule

Interactions. To more firmly establish the basis for the

underlying processes that control the course of the Al deposition,

(59) Parnis, M. J.; Mitchell, S. A.; Rayner, D. M.; Hackett, P. A. J. Phys. Chem.

1988, 92, 3869-3874.

9

J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002 5539

A R T I C L E S

Fisher et al.

data, as seen in Figures 8 and 13, show only small perturbations

of the spectral data with increasing Al deposition, which

suggests very weak interaction. In contrast, the XPS C 1s peak

associated with the terminal -CH₂OCH₃unit is not observable

for θ_AlJ 1.3 (Figure 9A), which suggests that the Al‚‚‚OCH₃

interaction perturbs the local electron density strongly to cause

the binding energy to shift completely into the region of the

main C 1s peak. Examination of the structural requirements of

the interaction geometry is helpful in exploring these issues.

Our theory calculations show that the minimum energy geometry

for the Al‚‚‚O(CH₃)R complex positions the Al atom on the

backside of the O atom away from the O-C bonds, as shown

below for an isolated -CH₂-O-CH₃unit. Simple examination

of the general surface structure of the H₃CO-terminated SAM,

as analyzed from space filling models, shows that unfavorable

steric interactions at the SAM surface provide a severe molecular

reconfiguration barrier to achieving this ideal “backside”

configuration, as illustrated in the adjacent cartoon.

this discrepancy is well outside of the inherent errors in the

calculations for the levels of theory considered, we conclude

that the Al interaction in the SAM is far weaker than the

minimum energy configuration predicted from theory, in line

with our analysis of the unfavorable steric interactions at the

SAM surface. On this basis, we view the interaction between

Al and -OCH₃more as a “solvation” rather than a complexation

with some directed interaction with a specific geometry. Given

the distribution of conformations of the OCH₃groups at the

SAM surface at ambient temperatures, especially in view of

the low rotational barriers around an alkyl C-O bond, relative

to an alkyl C-C bond, one can expect a wide distribution of

geometric configurations of Al relative to the surface atoms of

the SAM, analogous to a quasi-isotropic, weakly solvating layer

acting on average by mean-field interactions. We note that this

picture indicates that approach of an Al atom to the O-CH₃

bond, located at the outermost surface region, would seem not

to be sterically blocked. Therefore, we conclude that the lack

of Al insertion into this bond, a highly favorable process

thermodynamically for isolated model molecules, is due to an

activation barrier related to intrinsic electronic effects rather than

unfavorable surface configurations.⁵⁸Similarly, we conclude that

the observed lack of Al insertion into the C-O bond of the

-CH₂OH unit, located just beneath the OH SAM outer surface,

arises from intrinsic electronic and atomic reorganization effects

as well as unfavorable surface steric effects.

4.3.3. Comparisons of Reactivities of CH₃, OCH₃, CO₂CH₃,

OH, and CO₂H Groups. In this and previous reports, we have

characterized the reactivities of Al atoms with -CO₂CH₃,

-CO₂H, -CH₃, -OCH₃, and -OH moieties.^25-27Because the

first two groups can be viewed simplistically as combinations

of the latter moieties, along with the CdO unit, it is of interest

to check the self-consistency of the chemical reactivities toward

Al within the collection of these moieties. The -OH group

undergoes bond insertion with Al. Similarly, the main process

with -CO₂H appears to be O-H insertion.²⁷Neither the

-OCH₃nor the -CH₃group is degraded by Al. In the case of

the CO₂CH₃group, the OCH₃moiety remains intact, consistent

with the previous observation, while the CdO bond is com-

pletely degraded.^25,26Overall it appears that while CdO bond

attack is favorable relative to C-O, C-C, and C-H bonds, in

the presence of an O-H moiety, such as in the CO₂H group,

the CdO bond is less reactive than O-H. These comparisons

are meant to serve only as a qualitative guide. More detailed

bond reactivity correlations will require an expanded set of

molecular groups and careful analysis of experimental data.

While thermal energies might allow a small fraction of the

terminal -OCH₃groups to undergo conformational reconfigu-

ration for favorable Al interaction, it is clear that intermolecular

repulsion forces will prevent the vast majority from these

conformations. Thus one expects a substantial decrease in the

average stabilization energy per Al atom that can be achieved

in the SAM relative to the isolated configuration value. This

analysis fits qualitatively with the observed behavior of the IR

C-O stretching mode (Figures 12 and 13) that shows only a

minor shift and almost no change in line shape with deposition

of Al. These arguments apply to the OH SAM as well, although

the steric blocking should be less with H in place of OCH₃.

The weak interaction energies imply a fast dissociation of the

Al‚‚‚O bond in the complex.⁶⁰

Exploring this aspect further, the vibrational mode shifts for

the minimum energy Al‚‚‚O(CH₃)R complex were calculated

from theory [B3PW91/6-31G(d,p)]. The results of a detailed

analysis⁶¹for the isolated molecule CH₃(CH₂)₄OCH₃, setting

different chain and terminal group conformations and examining

all variations of modes involving some C-O stretch character,

show that the average shift for the C-O frequency is ∼100

cm^-1to lower frequencies upon interaction with the Al atom.

This shift is much larger, and even in a different direction, than

the observed shift, as can be seen in detail in Figure 13. Because

5. Conclusions and Future Work

The combination of the experimental data and theory calcula-

tions shows that the main difference between the interaction of

deposited Al atoms with an HO- and a H₃CO-terminated

hexadecanthiolate SAM on Au is that the OH group acts as an

efficient chemical trap for deposited Al atoms, preventing

penetration of the Al into the SAM matrix, while the OCH₃

group provides only very weak stabilization of the Al atoms,

analogous to solvation, thereby allowing Al penetration to

compete with overlayer film nucleation.

(60) The steric lowering of the ∼40 kJ/mol complex stabilization energies to

values approaching several kT in combination with activation barriers for

OH

I-O-H

OH

-C

OCH₃

IC-O

OCH₃

-C

insertion of ∼42-84 kJ/mol sets, k

< k

and k

, k

,

suggests that complexation could approach being a rapid preequilibrium

step.

In the case of the OH SAM, while chemical reaction of Al

with the OH groups is the dominant channel up to θ_Al≈ 1, the

reaction appears to cease, regardless of the Al coverage, before

(61) The details of more extensive calculations and analysis of interactions

involving other metals (e.g., Mg) and molecular groups, related to metal

atom deposition on SAMs, will be published elsewhere (Reinard, M. D.;

Cabarcos, O.; Allara, D. L., manuscript in preparation).

9

5540 J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002

Vapor-Deposited Aluminum Atoms

A R T I C L E S

all the OH groups are reacted, suggestive of steric blocking

effects. For 1.0 j θ_Alj 3-4, the deposited Al forms a

premetallic phase. Thereafter, a metallic overlayer is observed.

Experiments using isotopically labeled molecules with vibra-

tional spectroscopy and ultraviolet photoemission spectroscopy

would be useful in characterizing the character of the premetallic

phases, and such experiments are currently in progress.

nucleating overlayer, for example, the “wetting” of the metal

across the surface and the sizes of clusters formed. Variable

temperature, in situ atomic force microscopy would be useful

in characterizing these aspects, and such experiments are

underway in our laboratories.

Temperature-dependence studies, coupled with theory and

simulation studies, should prove very useful in determining the

detailed energetics of this system. Such studies, involving both

Al and related metal atoms (e.g., Mg), are in progress.

Overall, these studies reveal substantial details of the

interfaces that form from vapor deposition of Al atoms on

organic surfaces and provide a better understanding of the

specific mechanisms of chemical attack on the previously

studied H₃C-, H₃CO₂C-, and HO₂C-terminated SAMs. The work

makes clear the myriad of attainable chemical variations possible

with metallized organic systems and provides a basis for

designing and understanding the properties of many types of

Al metallized organic structures, particularly in applications such

as novel molecule-based electronic devices where extremely fine

control of the contact interfaces is critical.

In the case of the OCH₃SAM, the absence of a strong

chemical trap at the SAM surface opens a dynamic defect

penetration pathway to the SAM/Au interface where a stable

Al adlayer can form. The insertion of Al atoms at this interface

occurs with no appreciable effect on conformation or tilt angle

of the monolayer indicating the formation of a uniform adlayer.

Comparison with the previously studied case of a H₃C-

terminated SAM, for which the adlayer is shown to be complete

by θ_Al≈ 2.7 Al atoms/molecule (or ∼1:1 Al/Au for a Au(111)

surface) with a crossover thereafter to overlayer growth, leads

to the conclusion that the observed continuation of Al penetra-

tion well past this coverage for the OCH₃SAM is due to a

stronger surface stabilization of Al atoms by OCH₃than CH₃.

A combination of quantum theory calculations and molecular

conformation arguments shows that the Al interactions with the

surface atoms in the OCH₃SAM suffer strong steric effects

that reduce the average interaction energy from the ideal value

of ∼25 kJ/mol. The resulting weakened interactions can be

considered axially isotropic, analogous to a weak surface

solvation operating via mean field interactions, as opposed to

directed ones. This type of information is a useful complement

to a variety of gas-phase studies of the solvation interactions

of metal atoms and clusters with solvent molecules. In this

regard, one would expect a correlation between the interaction

SAM/metal interaction mechanism and the morphology of the

Acknowledgment. The authors acknowledge financial support

from the Office of Naval Research, the National Science

Foundation, and the National Institutes of Health. The authors

also acknowledge Alfred Miller and the XPS facility at Lehigh

University for their valuable cooperation in the XPS experi-

ments.

Supporting Information Available: Preparation and charac-

terization data for ¹⁸O-labeled compounds and ToF-SIMS data

for initial SAMs (PDF). This material is available free of charge

via the Internet at http://pubs.acs.org.

JA0123453

9

J. AM. CHEM. SOC. VOL. 124, NO. 19, 2002 5541

Article Doi

Bond insertion, complexation, and penetration pathways of vapor-deposited aluminum atoms with HO- and CH3O-terminated organic monolayers

DOI: 10.1021/ja0123453

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja0123453

Full text of DOI:10.1021/ja0123453

Products guided by the article

R&D Labs maybe for 2536-35-8

Relevant to this article

Hot Product