OFFICE OF NAVAL RESEARCH
R&T Code 413a001 ---01
"Thin Films from Solvated Metal Atoms and Metal-Metal Bonded Compounds"
Kenneth J. Klabunde
Department of Chemistry
Kansas State University
Manhattan, KS 66506
Reproduction in whole, or in part, is permitted for any purpose of the
United States Government
distribution is unlimited.
document has been approved for public release and sale;
i o n For
* Availability Codes
D T IC
%ECU.,ITY CLASSIFICATION OF THIS PAGE (*%on Data Ent .. reII
REPORT DOCUMENTATION4 PAGE
I. REPORT NUMBER
2. GOVT ACCESSION NO. 3. REC
C ATALOG NUMBER
TITLE (and Subtite)
"Thin Films from Solvated Metal Atoms and
Met-al-Metal Bonded Compounds"
Final Report August 1985-
TYPE OF REPORT a PERIOD COVERED
6. PERFORMING ORG. REPORT NUMBER
8. CONTRACT OR GRANT NUMBER(-)
Kenneth J. Klabunde
PERFORMING ORGANIZATION NAME AND ADDRESS
Kansas State University
Manhattan, Kansas 66506
1 1. CONTROLLING OFFICE NAME AND ADDRESS
10. PROGRAM ELEMEN-' -P.=,
& WORK UNIT NUMBERS
NUMBER OF PAGES
14. MONITORING AGENCY NAME & AODRESS(il different from Controlling Office)
I5. SECURITY CLASS. (of this report)
ECLASSI FICATION/ DOWN GRADING
800 N. Quincy Avenue
DISTRIBUTION STATEMENT (of thi. Report)
17. DISTRIBUTION STATEMENT (of the casract entered In Block 20. If dilloent from Reoiat)
I$. SUPPLEMENTARY NOTES
I9. KCEY WORDS (Continuo, on reverse ad. if necessary and Identify by block number)
metal atoms, solvated, nonaqueous colloids, metal particles, clustering,
living colloids, palladium, gold, indium, thin films, chemical liquid
ABSTRACT (Continue on reverse aide if nocoo.ctV and Identify by block number)
DD I JAN 7
EDITION OF INOV 65 IS OBSOLETE-
SECURITY CLASSIFICATION OF THIS PAGE (Whe.n Data Entered,
Metals such as Pd, Pt, Cu, Ag, Ga, In, Ge, Sn, and Pb are evaporated
under vacuum and the vapors (atoms) cocondensed at 77,K with excess organic
solvents. In this way solvated metal atoms are produced.. Upon warmup to
room temperature metal atom agglomeration occurs in certain solvents to
yield stable colloidal particles in solution. In many cases these are the
first examples of non-aqueous colloids of these metals, and they are very
novel in that they are free of contaminating reducing agents, halide ions,
etc., and they are living colloids--by removal of solvent metallic films can
be grown on various substrates under very mild conditions.
these colloidal particles and the films therefrom is an important part of
A second area is
compounds, eg. R Al-AlR2' , as possible new Chemical Vapor
for thin film production.
involving metal vapors, are underway. Compounds containing Al, Ga, In, and
The advantage of such material, on the microscopic
level, is that two metal atoms could be deposited at a time on a hot
structure/stoichiometry (when mixed metals are being codeposited).
the proposed synthesis of new metal-metal bonded
Unusual synthetic approaches, some
Brief Description of Results
Our colloid and surface work continued and we have reported on Au, Pd,
and In colloidal dispersions in organic solvents.
indium oxide were prepared and studied. We also found that stable Au and Pd
colloids can be prepared in liquid styrene, and this purple solution can be
controllably polymerized to metal doped polystyrene.
preparation of stable metal colloids such as these in a useful monomer
medium. Unusual stabilization mechanisms for these colloidal particles
Films of indium metal and
(1) the particles appear to scavenge electrons to become negatively
Electrophoresis studies, electron microscopy,
plasmon absorption spectroscopy, and conductivity studies have been valuable
in characterization of particles and films.
A second major emphasis this year has been the attempted synthesis of
Al-Al, Al-As, Ga-Ga, and Ga-As bonded organometallic compounds. We have met
with limited success. However, we have further elucidated the chemistry of
Al and Ga atoms, and As dimer. We have discovered a facile synthesis of
R Ga 2X
which may be an important discovery since it would allow a better
route o R 3Ga compounds for CVD processes.
must be important.
B. Recent Findings (July 1987 to July 1988)
The dispersion of metal vapor into fluorocarbon solvents has been of
interest in recent months. Metal atom clustering in cold fluorocarbons do
not stabilize colloidal suspensions at room temperature, and metal powder
fluorocarbon derived metal powder are "soluble" in organic solvents. A good
is gold powder derived from perfluoro-tri-n-butylamine.
extraction (or treatment of the filter cake) with acetone, red colloidal
suspensions are obtained.
The gold particles "dissolved" in this way are
smaller than those obtained directly from metal atoms dispersed in acetone.
How-ver, a very interesting phenomenon is
For example, fluorocarbon derived particles are
derived are 50-90A. Furthermore, the particles retain fluorocarbon;
according to X-ray fluorescence they are true gold but their surfaces
appear to be fluorocarbon coated, thus
(ability to be dissolved and small size).
10-40A, while acetone
their strange molecular behavior
We have also discovered that
methylmethacrylate can stabilize colloidal dispersions of Cu, Ag, Au, and
Pd. Polymerization of these colored solutions yields homogeneous metal
This procedure works extremely well and allows the
synthesis of various loadings of metal doped polymers.
C. A Listing of Technical Reports Submitted
1. S. T. Lin, M. T. Franklin, and K. J. Klabunde, "Non-Aqueous Colloidal
Clustering of Metal Atoms in Organic Media, 12," Langmuir, 2,
2. K. J. Klabunde, editor, "Thin Films From Free Atoms and Particles,"'
Academic Press, Orlando (1985).
3. K. J. Klabunde, "Introduction to Free Atoms and Particles," Chapter in 2
4. G. Nieman and K. J. Klabunde, "Clustering of Free Atoms and Particles:
Polymerization and Film Growth," Chapter in 2 above.
5. M. Franklin and K. J. Klabunde, "Living Colloidal Metal Particles from
Solvated Metal Atoms:
Clustering of Metal Atoms in Organic Media",
Sym. High Energy Processes in Organomet. Chem., K. Suslick, editor, ACS
Sym. No. 333, pg. 246 (1987).
Cardenas-Trivino, K. J. Klabunde, and B. Dale, "Living Colloidal
Palladium in Non-Aqueous Solvents.
Forming Properties. Clustering of Metal Atoms in Organic Media. 14",
Langmuir, 3, 986 (1987).
Formation, Stability and Film
Cardenas-Trivino, K. J. Klabunde, and B. Dale, "Thin Metallic Films
from Solvated Metal Atoms", Proceedings of SPIE, Vol. 821, 206 (1987).
8. K. Starowieyski and K. J. Klabunde, "Reactions of Vapors of Some Metals
and Metal Oxides with Organometallics of Main Group
Submitted to App, Organomet, Chem,
9. G. Cardenas-Trivino and K. J. Klabunde, "Characterization of Metallic
Thin Films Prepared by Chemical Liquid Deposition (CLD):
Colloidal Metal Particles in Non-Aqueous Solvents",
Chemical Society of Chile, in press.
Journal of the
G. Cardenas-Trivino and K. J. Klabunde, "Thermal Degradation Studies
and Scanning Electron Microscopy Studies of Pd Thin Films", XVIII
Latin-American Chemical Congress, Santiago, Chile, January 1988.
G. Cardenas-Trivino and K. J. Klabunde, "Synthesis, Stability, and
Structure of Metal Colloidal Dispersions in Non-Aqueous Solvents",
XVIII Latin-American Chemical Congress, Santiago, Chile, January 1988.
990 Langmuir, Vol. 3, No. 6, 1987
Cardenas-Trivino et al.
product. At higher temperatures products were evolved
that were probably formed from catalytic/pyrolytic de-
composition of acetone. In addition, ligand displacement
by excess (C6Hd)2P(CH5) yielded only acetone, and IR
studies suggest that the only displaceable organic material
is acetone itself. But note that it is quite strongly coor-
dinated, requiring a vacuum and warming for just partial
removal. A strong solvation mode is apparently important.
On the metal cluster surface a variety of binding schemes
may be operational, as suggested by Weinberg and Tem-
pleton2" for acetone on a Ru(001) surface:
are based on aqueous systems), it is clear that these neg-
atively charged particles will repel each other and therefore
aid their stabilization. l"
potentials are indicative of sub-
stantial electronic stabilization.
How is this negative charge acquired? One possibility
is that free radicals are involved, perhaps formed by py-
rolytic decomposition of small amounts of acetone on the
hot metal vaporization source or by reactions of acetone
with metal atoms. A number of radiolvsis studies of metal
colloids in water-acetone solutions indicate that organic
radicals do transfer electrons to the particles which act as
electron reservoirs (and can behave as catalysts for water
(CH3)2CtOH + (Ag),m - (CH3)2C'-OH + (Ag),-
CH3COCH3 + H+ + (Ag);-
If free radicals were involved in our system, the generation
of H+ in solution would be expected. However, we have
found no evidence of H+ in our solutions nor have we found
any radical recombination products that might be ex-
pected. Therefore, we do not believe free radicals are
important in the generation of negatively charged metal
particles in our system.
A second possibility is that the electron affinity of the
particles may allow them to acquire electrons from the
reaction vessel walls, electrodes, and solvent medium. Such
a process would help explain the need for a slow warmup
procedure in order to yield stable colloidal solutions since
scavenging of electrons may be a slow process.
Actually, this type of electrostatic charging of colloidal
particles is not uncommon. Oil droplets, for example,
scavange electrons from aqueous solution.1 4
If scavenging occurs during the warmup period, we
reasoned that by inserting a gold ground wire into the
solution during colloid formation some change would be
realized. Indeed, with this procedure the resulting Pd
particles became more highly negatively charged according
to electrophoresis studies (Table I). The next step was to
place a wire attached to the negative pole of a 12-V battery
into the solution during colloid formation. In this case
electrophoretic behavior changed markedly, and mea-
surements were impossible due to uncontrolled mixing. A
last case was to attach the wire to the positive pole of the
battery, and this again caused a significant change in the
behavior of the colloid. The migration rate was the highest
measured (Table I), and the colloid was very stable. Since
a circuit was not complete, either pole of the battery simply
served as a reservoir of electrons, and more electrons were
available, yielding more negatively charged particles.
On the basis of the above observations, we believe
particle stabilization occurs slowly during the warmup
period by steric effects (solvation) and by electronic effects,
where the growing particles develop and possess a suffi-
ciently high electron affinity that electron scavenging from
the reactor environment is possible. This scavenging can
be affected by the presence of electron sources, and elec-
trophoretic mobilities increased. Such experimental ma-
nipulations hold promise for controlling electrophoretic
mobilities and perhaps particle size.
Further support for this electronic stabilization mech-
anism is found in our studies of electrolyte additions. It
is known that electrolytes added to aqueous metal colloids
aid the breakdown of the charged double layer, which in
turn allows particle flocculation.'"-" Our studies with
electrolytes yielded similar results. The electrolytes with
As our Pd particles grow to hundreds of atoms, solvent
molecules would be incorporated within the particles and
on the outside. As growth continues, some solvent mole-
cules must be displaced by incoming atoms and smaller
metal particles. Eventually the particle growth stops. At
what point it stops (ultimate particle size) depends on the
initial metal concentration in the matrix and the matrix
Metal Concentration. Initial metal concentration can
affect colloid particle size in a kinetic way, since it is un-
likely particle growth is reversible under such conditions.21
Once a Pd-Pd bond is formed, it does not break. There-
fore, in a dilute solution of atoms, the frequency of en-
counters will be lower. As the metal atom-solvent matrix
warms and the atoms and the forming particle become
mobile, it is the number of encounters that occur during
the period before particle stabilization that is important.
And if metal concentration becomes too high, particle size
becomes too large, causing precipitation. Similar behavior
has been encountered for gold colloids in acetone.21 In-
terestingly, however, gold particle size could be more easily
controlled by concentration effects.21 With palladium we
invariably obtained particle sizes of 6-12 nm. Low con-
centrations of Pd still yielded 6-8-nm particles, and high
concentrations of Pd yielded 8-12-nm particles plus much
larger particles that precipitated. Thus, there is a distinct
preference for an 8-nm average particle size for Pd in
acetone as well as for Pd in ethanol. We do not fully
understand this selectivity yet, although particle stabili-
zation must be the key, as discussed below,
Particle Stabilization. We believe particle growth
stops because of two factors. The first comes under the
heading of steric stabilization.' Solvent molecules must
be displaced and reordered on the surface of a Pd cluster
if another cluster is to chemically bind to it. As the par-
ticles (clusters) become more massive the kinetic energy
goes down, and perhaps the energy requirement for solvent
displacement/reordering becomes large compared to the
kinetic energy of the sluggish larger particles.
A second mode of stabilization is electronic in nature.
Electrophoresis experiments clearly show that the Pd
particles bear negative charge. Although it is difficult to
determine accurately the number of negative charges each
particle possesses (formulas derived for such calculations
(20) Templeton, M. K.; Weinberg, W. H. J. Am. Chem. Soc. 1985, 07,
(21) Franklin, M. F.; Klabunde, K. J. High Energy Processes in Or.
ganometallic Chemistry; Suslick, K. S., Ed.; ACS Symposium Series 333;
American Chemical Society: Washington, DC, 1987; pp 246-259.
(22) Henglein, A. J. Am. Chem. Soc. 1979.83, 2209-2216.
(23) Henglein, A.; Lillie, J. J. Am. Chem. Soc. 1981, 103, 1059-1066.
Living Colloidal Palladium in Nonaqueous Solvents
Langmuir, Vol. 3, No. 6, 1987 991
the more highly charged cations caused flocculation more
qtickly (AIl3 > Ca2+ > Na+). This is a classic case showing
not only the existence of charged colloidal particles but
also that a charged double layer must exist."
Living Colloids - Films. From Table II it is evident
that substantial portions of organic residue remain in the
films after solvent stripping at room temperature. We
found that the films were susceptible to oxidation, as might
be expected, and oxygen (by difference) ranged as high as
25%. If care was taken to prevent oxidation, an empirical
formula of about Pd5C-H202 was determined. An average
of all determinations indicated PdgCH7O11. After treat-
ment at 500 *C, causing the evolution of some organic
material, an average empirical formula of PdC103 was
under vacuum by removal of the liquid N2 filled Dewar for 1.5
Upon meltdown a black solution was obtained. After addition
of nitrogen the solution was allowed to warm for another 0.5 h
to room temperature. The solution was siphoned out under N2
into Schlenk ware. Based on Pd evaporated and acetone inlet
the solution molarity could be calculated.
Effects of a Ground Wire and a Battery-Attached Wire.
Several experimenth were carried out where a gold wire was
connected to an electrode inside the reactor so that it reached
the bottom of the reactor. A copper wire was attached to the upper
part of the electrode external to the vacuum chamber. This wire
was either grounded or attached to the negative or positive pole
of the 12-V storage battery. Colloidal solutions obtained by using
these modifications did not show any marked changes in stability,
but electrophoretic mobilities increased.
Electrophoresis Experiments. The electrophoresis exper-
iments were carried out by using a glass U-tube of 11.0 cm with
a stopcock on the base to connect a perpendicular glass tube (13
cm long X 35 cm high)."'-m Platinum electrodes were attached
to the top of the U-tube and through a ground glass joint to the
pole of a 12-V battery. The acetone was placed in the U-tube,
and then the colloid solution was added slowly thragi. ,he side
tube. The migration rate was determined on the basis of the
aNerage of the displacement in each side of the U-tube. A typical
experiment was carried out for a period of 3 h at 25 9C.
Electrolyte Additions. A study of flocculation times was
carried out by using a 0.010 M Na! solution. In a test tube 2 mL
of colloidal solution (0.0175 M) and 2 mL of Nal were added at
room temperature (25 *C). After 5 min flocculation of the colloid
A solution of 0.010 M Cal2 in acetone was also prepared. By
use of the same ratio as before, flocculation of the colloid began
after 3 min. Finally, a 0.010 M AIBr3 acetone was prepared.
Addition to the colloid in the same amount as before induced
flocculation after 1 min at room temperature. Complete floccu-
lation was observed after 10, 8, and 7 min, respectively.
In other experiments water was added to the colloid solution,
and after 120 h flocculation was observed.
GC-MS Experiments. GC-MS pyrolysis was carried out by
using a Porapak Q 6-ft column (flow rate 35 mL/min) attached
to a Finnigin 4000 quadrupole GC-MS. The sample was placed
in a stainless steel tube 10 cm long connected to a four-way valve.
One of the outlets was attached to a Porapak Q column interfaced
with the M.S. The stainless steel tube containing a portion of
Pd colloid film (stage Ill was placed in a furnace connected to
a Variac provided with a digital quartz pyrometer to measure the
temperature. Three pyrolyses were performed at 100, 200, and
350 *C with the Pd-acetone film (from colloid, 0.0521 M).
Addition of (C6H5)2P(C2H5). A Pd film was prepared by
evaporating the solvent from a 0.035 M colloid solution. A 25-mg
sample of the film was treated with 1.5 mL of (C6HS)2 P(C2H5)
(5.6 mmol) under a nitrogen atmosphere. After 48 h at room
temperature with stirring, the dark solution became lighter. The
volatiles were pumped out through 263 and 77 K traps. The 77
K trap contained only acetone (0.39 mg or 6.8 X 10-3 MMol),
identified by gas-phase IR.
SEM and TEM Studies. Electron micrographs were obtained
on a Jeol, Temscan-100 CXIl combined electron microscope and
a Hitachi HV-IIB (TEM) operated at 2 X I0 magnification. The
specimens for TEM were obtained by placing a drop of the colloid
solution on a copper grid coated by a carbon film. The samples
for SEM were placed between two copper grids one of which was
coated by a carbon film.
Resistity Studies. Films of different thickness (02.8-65 um)
were prepared by dripping the colloidal solutions onto a glass plate
edged with silicon rubber adhesive resin. The acetone was allowed
aese resnTe acetone slo
o evaporate. Resistivities were measured by scraping the silicon
rubber away from the edges of the film, which was then trimmed
to rectangular shape. iL was men connectea to electrodes on each
end by vapor deposition of an opaque film of aluminum or copper.
To get a reliable contact on aluminum, it was necessary to apply
a spot of silver paint over the aluminum. This was not necessary
* .- "Z
Earlier discussion suggested that while in solution the
colloidal particles are solvated by acetone, and other or-
ganic fragments were not detected. However, upon solvent
it is obvious from the empirical formulas that
acetone must be breaking up, accompanied by some oxi-
dation. Some acetone is still present since it is the main
volatile product evolved at 300 *C (Table III) and the only
volatile product displaced by (C6HS)2P(C2H5). Since the
remaining fragments must be very rich in carbon and ox-
ygen, the formation of palladium carbides and palladium
oxides is likely.
Electron microscopy studies show that the individual
colloidal particles are spherical and have a tendency to link
together in chains. The initial film appears to be made
up of a network of Pd particle chains (Figure 2). Heating
the film causes these chains to collapse to a more uniform
film (Figure 3).
Resistivities of these films are of interest. Table IV lists
values determined for 1 cm2 films of varying thickness
(0.2-60 jim). The initial films are conductive and increase
in conductivity after heating. They behave more like
semiconductors than pure metals, and actually their re-
sistivities are similar to those of doped organic polymers.'
Palladium atoms dispersed in excess acetone (or other
solvents) begin to cluster upon warming. The properties
of the resultant colloidal particles depend slightly on initial
metal concentrations, warmup procedures, and the
availability of electrons. During colloid formation the
particles are stabilized by solvation effects and by elec-
tronic effects due to electron scavenging (the Pd particles
behave as electron sinks). Upon solvent removal, films of
intertwined chains of spherical Pd colloidal particles (still
containing organic residues) are formed. Heating the film
causes these chains to collapse to a uniform film with a
decrease in resistivity. Some organic residue remains in
atom reactor has been described previously.iihg As a typical
example, a W-A203 crucible was charged with 0.80 g of Pd metal
(one piece). Acetone (300 mL, dried over K2CO3) was placed in
a ligand inlet tube and freeze-pump-thaw-degassed with several
cycles. The reactor was pumped down to 1 X 10' Torr while the
crucible was warmed to red heat. A liquid N2 filled Dewar was
placed around the vessel, and Pd (0.5 g) and acetone (189 g) were
codeposited over a 1.0-h period. The matrix was dark-brown at
the end of the deposition. The matrix was allowed to warm slowly
(24) Wegner, G. Angew. Chem., Int. Ed. Engl. 1981, 20, 361-381.
(25) Klabunde. K. J.; Timms. P. L. Ittel, S.; Skell, P. S. Jnorg. Synth.
1979, 19, 59-86.
(26) Shaw, D. J. Electrophoresis; Academic: New York, 1969.
with copper electrodes. The resistance of each sample was
measured with a Keitley 178 Model digital multimeter. the vapor
depositions were carried out with a Veeco Model VS-90 metal
evaporator. the values of thickness and resistance are summarized
in Table IV.
Solubility Studies. The solubility of the Pd-acetone film
(0.0236 M) was tested with the following solvents: acetone,
ethanol, THF, DMSO, benzene, toluene, and pentane. The films
were completely insoluble after 24 h of contact with stirring at
Infrared Studies. Infrared spectra were recorded on a
Perkin-Elmer PE-1330 infrared spectrometer. IR studies of the
metal films using either KBr pellets or Fluorolube yielded only
evidence for Pc-_ (2980 cm-1) and PO (1740 cm-) showing the same
shape as the acetone standard.
Acknowledgment. The support of the Office of Naval
Research is acknowledged with gratitude. We also thank
Rarh is akno
Matthew T. Franklin for helpful discussions and Larry L.
Seib for assistance with the SEM-TEM experiments. Also
we want to thank Dr. Ileana Nieves for her assistance in
obtaining spectra and Thomas J. Groshens for assistance
with the mass spectrometer.