5*;
L35?

t;

:z
15%
.i

€
5

 

6'3

i;

. 5:11.5"3‘9;
iii»...i3l$
“Ronni! .
oJ. in}?
. I .‘x Vt
I.

 

a.
i?! .
. \.:.Ixo.lk
$1,.

.I.v....o I
O ,5.

 

._.T'TT","""

 

'v' v" vv I‘o

 

 

 

v-I.
y 1.;v»l../l.

 

     

   
   

 

 

Wt; ., - '... H.....|...;.w..ut......? .‘ou.v.yfln....{t. ’03...ka

. ‘

IL}... nest»... v.5 . ‘ : . r .2: .7

u ntl .
‘ V 51.3.... .v .. .
it...

 

 

 

WHHHUNN(”HMWHIIHIHIHHWINIHIUHIHIHII

300885 3321

Thisistocertifythatthe

Nansen-’6’ MewdmuonfiwP am 507 #5th

KMO *LJh T50]

has been accepted towards fulfillment
of the requirements for

Ph. 0. (min Ckemtd‘gfi

, //}er/Z
<fl (17¢

 

 

 

//I

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

N AN OSCALE METAL AND ALLOY PARTICLES BY HOMOGENEOUS
REDUCTION WITH ALKALIDES OR ELECTRIDES AND REDUCTION IN
ZEOLITE PORES

By

Kuo-Lih Tsai

A DISSERTATION
Submitted to
Michigan State University

in partial fulfillment of the requirements
for the degree of

DOCTOR OFPHILOSOPHY
Department of Chemistry

1991

ABSTRACT

NANOSCALE METAL AND ALLOY PARTICLES BY HOMOGENEOUS
REDUCTION WITH ALKALIDES OR ELECTRIDES AND REDUCTION IN
ZEOLITE PORES

By

Kuo-Lih Tsai

A new method for the preparation of small metal/oxidized
metal or alloy particles and reduction in zeolite pores is described,
that utilizes homogeneous reduction of metal salts by dissolved
alkalides or electrides in an aprotic solvent such as dimethyl ether
or tetrahydrofuran. Soluble compounds of transition metals and
post-transition metals in dimethyl ether or tetrahydrofuran are
rapidly reduced at -30 °C by dissolved alkalides or electrides to
produce metal particles with crystallite sizes from < 3 to 15 nm.
The average particle size was estimated from the line broadening
of powder X-ray diffraction. Particle size distributions were
determined by counting the particles on electron micrographs
obtained by transmission electron microscopy. Salts of Au, Pt, Cu,
Te, Fe, and .Ta formed metallic particles with little or no oxidation
even when washed with methanol or liquid ammonia. Reduction
of salts of Ni, Fe, Zn, Ga, Si, Mo, W, In, Sn, and Sb yields surface
oxidation over a metallic core. This method is also applicable to

the formation of finely divided metals on oxide supports.

Alloys or intermetallic compounds formed when two
different metal salts were used, as indicated by X-ray
photoelectron spectroscopy. Confirmation by electron diffraction
was made in the case of air-stable samples. Reduction of salts of
Au-Cu, Au-Zn, Cu-Te, Zn-Te, Au-Ti, Fe-Ta formed particles that
contained intermetallic compounds.

Stoichiometric amounts of the alkalide or electride were
used; these were prepared separately or in situ. FT-IR
spectroscopy was used to demonstrate that the presence of
organic complexants and solvents often resulted in the inclusion of
organic decomposition products due to the high reactivity of
nanoscale metal particles. Elemental analysis showed that both
carbon and hydrogen were less than one percent in atomic
concentration on the surface of gold.

Electrides, the most powerful reducing agents, were used to
reduce Cu2+, Pd2+, Ag+ in zeolite-Y. X-ray photoelectron
spectroscopy data show that these cations were completely
reduced to the zero oxidation state at -30 °C. Cu and Pd remain
trapped in the interior and an alkali metal cation balances the
counter charge upon reduction. In contrast, Ag formed 40 A
diameter particles uniformly on the external surface. This method
should also be applicable to the synthesis of organometallic

compounds.

1:0 my parzuts aub mum

iv

ACKNOWLEDGEMENTS

I would like to express my appreciation to Dr. James L.
Dye for his continual guidance and support throughout this
work. I would also like to thank Dr. Kim Dunbar, Dr. Mercoury
Kanatzidis and Dr. James Harrison for their helpful discussions
and advice.

Appreciation is also extended to all of the members of the
Dye research group with whom I have worked, in particular
Drs. Rui Huang, Evy Jackson, Jineun Kim, Kevin Moeggenborg,
and Judith Eglin. I would like to thank the Master Glassblowers
at Michigan State University, Keki Mistry, Manfred Langer, and
Scott Bancroft, for the numerous pieces of glassware that they
have designed, built, and repaired for me. I am grateful for the
help from Dr. Karren L. Klomparens with the transmission
electron microscopy, Dr. Kevin J. Hook with the X-ray
photoelectron spectroscopy and Dr. Debbie Duxbury with the
powder X-ray diffractometry.

I am grateful for finacial support from the National
Science Foundation (Solid State Chemistry Grants Nos. 87-
14751 and 90-17292) and the Center for Fundamental
Materials Research at Michigan State University and the 1989
Graham summer fellowship, the 1990 Yates summer fellowship

from the Chemistry Department, Michigan State University.

TABLE OF CONTENTS

LIST OF TABLES
LIST OF FIGURES

Chapter 1

Chapter 2

Chapter 3

Introduction
A. Alkalides and Electrides
B. Small Metal Particles
C. Theory
(1) X-ray' Photoelectron Spectroscopy
1. General background
2. Nomenclature
3. Chemical shifts
(II) Powder X-ray Diffraction Line Broadening
(III) Electron Diffraction
D. Objective

EXPERIMENTAL METHODS

A. Synthesis

B. X-ray Photoelectron Spectroscopy
C. Powder X-ray Diffraction

D. Transmission Electron Microscopy

SINGLE ELEMENT REDUCTION
A. Reduction with Alkalides or Electrides
B. Reducing Agents
C. Results

(1) Gold

(11) Copper

(III) Tellurium

(IV) Nickel

(V) Antimony

(VI) Gallium

(VII) Zinc

(VIII) Molybdenum

(IX) Tin
D. Summary

vi

Chapter 4

Chapter 5

Chapter 6

APPENDIX

BINARY ELEMENT REDUCTION
A. Introduction
B. Experimental
C. Results
(1) Au-Zn
(II) Au-Cu
(III) Tellurides
(IV) Ti-Au
(V) Ta-Fe
D. Summary

REDUCTION in ZEOLITE PORES
A. Introduction

B. Experimental

C. Results

D. Summary

CONCLUSIONS and SUGGESTIONS for FUTURE WORK

REFERENCES

vii

63
63
63
64
65
68
70
71
73
76

77
77
82
83
92
93

99
104

Table 1-1.

Table 1-2.

Table 3-1.

Table 3-2.

Table 3-3.

Table 3-4.

Table 4-1.

Table 4-2

LIST OF TABLES

X-ray and spectroscopic notation

Spin-orbit splitting parameters

Electron diffraction of Au from reduction of AuC13
Electron diffraction of Cu from reduction of CuClz

Electron diffraction of Te from reduction of TeBr4

Part of the periodic table showing the elements

studied to date. Identification: 1, washable metals;

2, surface-oxidized; 3, nonwashable; 4, reoxidized.

See the text for more detailed definition

Electron diffraction of AuZn from co-reduction of
AuCl3+Zn12

Electron diffraction of AuCu from co-reduction of
AuCl3+CuC12

viii

35

42

51

62

67

69

Figure

Figure

Figure

Figure

Figure
Figure

Figure

Figure

Figure

Figure

Figure

Figure

1-1.

1-3.

1-4

1-5.

1-6.

2-1.

3-1.

LIST OF FIGURES

Figure 1-1. Representative complexants :
15-Crown-5 (IUPAC name: 1,4,7,10,13-
pentaoxacyclopentadecane) and Cryptand [222]
(IUPAC name: 4,7,13,16,21,24-Hexaoxa-l,10-
diazabicyclo [8.8.8]hexacosane) 1

Diagram of the X-ray photoelectron process 5

Diagram of the fluorescent X-ray photoprocess
and the Auger process 6

Two-dimensional zinc chemical state plot for XPS
(from tabulation by C. D. Wagner, Handbook of X-
ray Photoelectron Spectroscopy; Perkin-Elmer;

1979; p84) 12
Diagram of X-ray diffraction l4
Diagram of electron diffraction 16

Kontes double tube("H-cell") use to prepare metal

and alloy particles 20

XPS spectrum of Au from the product of reduction
of AuC13 3 1

XRD lines( 111) used to determine average particle

sizes for An. 32
Particle size distributions for Au 33
Electron micrograph of Au

(mag=lO0,000X, 1,000 A/cm) 34
Selected area electron diffraction of Au.

(camera length=83 cm) 34

ix

 

 

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

3-7.

3-8.

3-9.

3-10.

3-11.

3-12.

3-13.

3-14.

3-15.

Figure3-16.

Figure

Figure

Figure

Figure

3-17.

3-18.

3-19.

3-20.

Electron micrograph of Au particles on
A1203(mag=190,000x, 550 A/cm) 36

Optical absorption spectra of an Au colloidal
solution 37

XRD spectrum of the products from the reduction
of AuCl3. (A=6.5x10‘4 M, B=1.6x10'4 M) 3 8

FT -IR spectra of washed Au particles 40

XRD lines(111) used to determine average
particle sizes for Cu 41

Selected area electron diffraction of Cu

(camera length=140 cm) 42
Electron micrograph of Cu

(mag=320,000X, 310 A/cm) 44
Particle size distribution of Cu 44

Energy dispersive spectra of Cu particles on
Ni grid 45

XPS spectrum of Cu (A before washing, B
exposed to air for 24 hours, and C after washing) 46

XPS spectrum of Te (A before washing, B exposed
to air for 3 min, and C after washing) 48

XRD lines(100) used to determine average
particle sizes for Te 49

Electron micrograph of Te
(mag=140,000X, 700 A/cm) 50

Selected area electron diffraction of Te
(camera length=140 cm) 50

XPS spectrum of Ni.(A) before washing (B) oxidised
in the air for 24 h (C) washed by MeOH 5 3

X

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

3-21.

3-22.

3-23.

3-24.

3-25.

4-1.

4-2.

4-3 .

5-1.

XPS spectrum of Sb. (A) before washing (B) washed
by MeOH (C) sputtered by Ar ions 54

XPS spectrum of L 3M45M4 5 Auger peaks of Ga.

(A) reaction products (B) sputtered by Ar ions
for 2 min (C) sputtered by Ar ions for 5 min 55

XPS specrum of L 3M45M45 Auger peaks of Zn.
(A) reaction products (B) sputtered by Ar ions
for 2 min (C) sputtered by Ar ions for 4 min 57

XPS specrum of 3d peaks of Mo. (A) reaction
products (B) sputtered by Ar ions for 10 min
(C) sputtered by Ar ions for 30 min 58

XPS specrum of 3d peaks of Sn. (A) reaction products
(B)exposed to the air for 30 seconds 59

XPS spectrum for the 4f5/2 and 4f7/2 levels

of Au:(A) Au particles from AuCl3 reduction;
(B)AuZn particles formed from reduction of

a mixture of AuCl3 and excess ZnIz;

(C) AuZn + Au particles formed from reduction

of a mixture of AuCl3 and Zn12 with the former

in excess 66

XPS of the reduction product of a mixture of
TiCl4 and AuCl3 in MezO (top) and after
washing with air-free methanol (bottom) 72

XPS results for the Ta-Fe system; product washed
with anhydrous NH3: (A) Fe XPS from 2:1 FeCl3
to TaC15; (B) Fe XPS from FeCl3 alone; (C) Ta XPS
from 2:1 FeCl3 to TaC15; (D) Ta XPS from

TaC15 alone 75
Structure of sodalite cage 79
Structure of zeolite Y 80

xi

Figure 5-3.

Figure 5-4.

Figure 5-5.

Figure 5-6.

Figure 5-7

XPS spectra of Pd in zeolite Y before reduction
3d5/2=337.5 eV, 3d3/2=343.0 eV (top) and after
reduction 3d5/2=335.2 eV, 3d3/2=340.4 eV
(bottom). 84

The XPS survey scans of (A) plain zeolite Y (B) after
NadeCl4 ion exchange~( new peaks of Pd 3d5/2 at
337 eV and 3d3/2 at 343 eV appear and Na ls peak
at 1071 eV decreases) (C) after reduction by ,
K+(15C5)2e',(new peaks of K 2p appear showing
potassium cations balance the counter charge) 86

XPS spectra of Cu in zeolite Y before reduction
(2p3/2= 934.35 eV, 2p1/2=954.15 eV top), after
reduction (2p3/2=933.7 eV, 2p1/2=953.8 eV middle),
and oxidized in air (2p3/2=935.8 eV, 2p1/2=955.9 eV
bottom) 88

XPS spectra of Ag in zeolite Y before reduction

3d5/2=369.5 eV, 3d3/2=375.6 eV (top) and after

reduction 3d5/2=368.7 eV, 3d3/2=374.7 eV (bottom)
90

The XPS survey scans of (A) plain zeolite Y (B) after
AgNO3 ion exchange ( new peaks of Ag 3d5/2 at 369
eV and 3d3/2 at 375 eV appear and Na 1s peak at
1071 eV disappear) (C) after reduction by ,
K+(15C5)2e',(the peak intensities of Ag show surface
enrichment and new peaks of K 2p appear showing
potassium cations balance the counter charge) 91

xii

Chapter 1 I NTRO DUCTIO N

A. Alkalides and Electrides

Alkali metals were found to dissolve in liquid ammonia by
Weyl in 1864.(l) Since then. the study of metal ammonia solutions
has become an active research area. Two major species present in
the blue ammonia solutions are the metal cations (M") and the
solvated electron (e'solv)- Crown ethers and cryptands were
synthesized by Pedersen and Lehn respectively in the l960's.(2. 3. 4)
Crown ethers are cyclic polyethers and cryptands are bicyclic
polyethers that readily complex metal cations: examples are shown

in Figure 1-1.

(0/). (We.
<_O 0‘7
\__/

lS—Crown-S (15C5) Cryptand [2221 (C222)

Figure l—l. Representative complexants : lS-Crown-S (IUPAC name:
1.4.7.10. l3-pentaoxacycIopentadecane) and Cryptand
[222] (IUPAC name: 4.7.13.16.21.24-Hexaoxa-1JO-

diazabicyclo [8.8.8]hexacosane).

2

in the early 1970's. Dye and co-workers combined alkali
metals with crown ethers and cryptands in amine or ether solvents
to produce a new class of compounds. which they called alkalides.
Alkalides are salts in which the cation is a complexed alkali metal
cation. and the anion is that of either the same or a different alkali
metal. The first alkalide. Na*(C222)Na’. was reported in 1974(5)

By changing the ratio of metal and complexant. Dye and co-
workers synthesized another new class of compounds called
electrides. Electrides have the same complexed cation as alkalides
but the anion is replaced by a trapped electron. The first crystalline
electride. Cs+(18C6)2e’. was reported in 1983(6) So far. more than
40 alkalides and 10 electrides have been synthesized by using
various alkali metals and complexants.

These compounds crystallize from solution to yield shiny
bronze-colored crystals (alkalides) or black crystals (electrides)
which are all reactive towards air and moisture and thermally
unstable at room temperature. Because of their unusual nature and
properties. they have been studied by a variety of experimental
methods. These include optical studies.(7) nuclear magnetic
resonance spectroscopy.(8) electron paramagnetic resonance
spectroscopy.(9) magnetic susceptibility.(10) single crystal X-ray
crystallography.(l 1) powder X-ray diffraction.(12) photo emission
spectroscopy.( 13) and pressed powder electrical conductivity.(l4. 15)

Because alkalides and electrides are air sensitive and thermally
unstable. their synthesis must be done at low temperatures in the
absence of air. moisture and other reducible substances. The

synthesis of alkalides and electrides via vacuum line techniques has

3

been described in detail elsewhere so only a brief description will be
given here.(l6. 17)

The apparatus usually used to synthesize these compounds is
called a “K cell”.( 16) To synthesize either an alkalide or an electride.
the reactants are placed in a reaction vessel in a helium filled inert
atmosphere dry box. The complexant is added to one chamber of the
cell through one side-arm and the metal is added to another side-
arm in an opened glass ampoule. The side-arms are then closed with
an Ultra-Torr connector and a sealed glass tube. After removal from
the dry box the metal in the ampoule is distilled into one chamber of
the cell and the side-arms are sealed off. This is accomplished by
heating the ampoule with a torch which produces a thin film of metal
in the chamber (Note that this cannot be done with lithium).

A solvent such as methylamine or dimethyl ether is distilled
into the complexant side of the cell. The solvent is used to dissolve
both the complexant and the metal by repeatedly washing from one
chamber of the cell to the other. Next. most of the first solvent is
removed and a second less polar solvent is added to precipitate the
product. The solution is then allowed to stand for several hours.
after which the precipitate in the chamber-is washed several times
with the second solvent. 'After final removal of the second solvent.
the polycrystalline precipitate is poured into the ”fingers“ at the top
of the cell and vacuum-sealed to prevent decomposition. The
compounds are stored in a liquid nitrogen dewar or a -80 'C freezer

until needed.

4
B. Small Metal Particles

The preparation of nanoscale particles (1-20 nm diameter) as
colloids or aggregates is a well-developed field that involves a
variety of chemical and physical techniques.(18. 19) Small noble-
metal particles are commonly made by mild reduction.(20. 21. 22)
Rieke and co-workers and others have reduced salts of more active
metals in ethereal or hydrocarbon solvents. either heterogeneously
with alkali metals (slow) or homogeneously with aromatic radical
anions such as naphthalene (fast).(23. 24. 25. 26. 27) The products of
such reactions are highly active metal powders. Other methods such
as pyrolysis of precursors.(28)evaporation of metals.(29) matrix
isolation (solvated metal atom dispersion).(30) and sol-gel processes
(31) have also been used to prepare small metal particles. A more
recent development is reduction with alkali metal
organoborohydride solutions. such as NaB(Et)3H. that has been shown
to yield both single metals and alloys of the iron-group elements and

the noble metals.(32)

C. Theory

(1). X-ray Photoelectron Spectroscopy

1. General background
X-ray Photoelectron Spectroscopy (XPS). which is also
commonly termed electron spectroscopy for chemical analysis (ESCA)

is one of a number of surface analysis techniques. in addition to its

5

surface sensitivity. XPS is capable of providing chemical information
such as the oxidation state and the nature of the chemical bonding. as
well as elemental composition. Therefore. it is a very good method
for the characterization of nanoscale particles.

XPS involves the energy analysis of electrons ejected from a
surface under bombardment by soft X-rays. Mg Kc: X-rays (1253.6
eV) or Al K0: X-rays (1486.6 eV) are ordinarily used. These photons
have limited penetrating power in a solid. of the order of 1-10
micrometers. The photoelectron ejection process occurs when a core
level electron absorbs a photon of energy greater than its binding
energy. When this occurs. the electron is ejected from the atom with
an energy characteristic of the exciting photon and the initial core

level binding energy as illustrated in Figure 1-2.

4 4i3p}M

I 3s

hull. _
K it ,2. L

131K

 

Figure 1-2. Diagram of the X-ray photoelectron process.

The emitted electrons have kinetic energies given by
KE- hv - BE - 4:5
in which [(13 is the kinetic energy of the photoelectron.
ho is the X-ray photon energy.

6

BE is the photoelectron binding energy.
4’s is the spectrometer work function.

When a sample is illuminated by an intense source of photons
of a single well defined energy the resultant photoelectrons can be
resolved into energy peaks that are characteristic of the emitting
atoms. In addition to the photoelectrons emitted in the photoelectric
process. Auger electrons are emitted due to relaxation of the
energetic ions that remain after photoemission. Auger electron
emission occurs roughly 10"” seconds after the photoelectric event.
More than 997. of the energy of relaxation shown in Figure 1-3 is
released through Auger processes. Fluorescent X-ray photon
emission is another process in this energy range. but occurs less than

one percent of the time.

_4_
IL

4 4 3p L
Tr— 3s}M

_l___L 2p}L IL

‘_ 25
ls |K

‘
=I‘

ht)
X- ray Emission Auger EmiSSiOl‘l

Figure 1-3. Diagram of the fluorescent X-ray photoprocess and
the Auger process.

7

1n the Auger process. an outer electron falls into the inner
orbital vacancy. and a second electron is emitted. carrying off the
excess energy. The Auger electron kinetic energy is equal to the
difference between the energy of the initial ion and the doubly-
charged final ion. This process is independent of the mode of the
initial ionization. Thus photoionization often leads to two emitted
electrons. a photoelectron and’an Auger electron.

The path length of the photons in the material is of the order of
micrometers. but that of the electrons is of the order of tens of
Angstroms. Thus. while ionization occurs to a depth of a few
micrometers. only those electrons that originate within tens of
Angstroms of the solid surface can leave the surface without energy
loss. It is these electrons that produce the peaks in the spectra and
are most useful. Those that undergo loss processes before emerging

contribute to the background.

2. Nomenclature

Since an electron is a charged particle. its orbital motion around
a nucleus induces a magnetic field whose intensity and direction
depend on the electron velocity and on the radius of the orbit.
Quantum mechanically. this is characterized by the orbital angular
momentum operator. The appropriate quantum number I can take
the values 0.1. 2. 3. ...... n-l in which n is the principal quantum
number. Another property of an orbiting electron is its intrinsic

spin. lts spin quantum number s can take either of the values :1: 1/2.

8

The total electronic angular momentum is a combination of the
orbital and spin angular momenta.

The total angular momentum of a single isolated atom is
obtained by summing vectorially the individual electronic spin and
angular momenta. For a particular electron. the total angular
momentum is characterized by the quantum number 1. where l-l+s.
The value of i can be 1/2. 3/2. 5/2 ..... etc. The total angular
momentum for the whole atom is the summation for all electrons. J
-2‘. 1,. This process is known as j-lcoupling and occurs for inner core
electrons in atoms.

Under l-icoupling the nomenclature is based on the principal
quantum number n and the electronic quantum numbers I and I. In
the X-ray notation. states with ml. 2. 3. 4.... are designated K. L. M.
N.... respectively. while states with various combinations of 1-0. 1. 2.
3.... and 1-1/2. 3/2. 5/2. 7/2.... are given conventional suffixes.l. 2. 3.
4 .....

A spectroscopic nomenclature has been developed that is
equivalent to that used with X-rays. The principal quantum number
appears first. then the states with l-O. l. 2. 3.... are designated 5. p. d.
f. respectively and the [values are appended as suffixes. Thus the
state written L2 in the X-ray notation. in which n-2.(-1 and 1-1/2.
would be written 2p“; in spectroscopic notation. The relation
between X-ray and spectroscopic notation is summarized in
Tablel-l.

9

Table 1-1. X-ray and spectroscopic notation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

n t I X-ray suffix X-ray level spectroscopic level

1 0 1/2 1 K 151/;

2 0 1/2 1 L1 251/2

2 1 1/2 2 L2 291/2 __
2 1 3/2 3 L3 293/2 .
3 0 1/2 1 MI 351/2

3 1 1/2 2 M2 SpL/z

3 1 3/2 3 M3 393/;

3 2 3/2 4 M4 303/; .
3 2 5/2 5 M5 305/;

etc. etc. etc. etc. etc. etc.

 

 

 

Spin-orbit coupling results in splitting of the atomic p. d. f

energy levels into mm and p3/2. (13/2 and ds/z. and Ba and fig

components. The relative intensities of the doublet peaks are given

by the ratio of their respective degeneracies (2]+1). The area ratio

and designations (0(5) of spin-orbit doublets are given in Table 1-2.

Table 1-2. Spin-orbit splitting parameters

 

 

 

 

 

 

 

 

Subshell jvalue Area ratio
5 1/2 ---
p 1/2 3/2 1:2
d 3/2 5/2 2:3
f 5/2 7/2 3:4

 

 

10

The X-ray notation is almost always used for Auger transitions.
so that. for example. in 1—] coupling there would be six predicted KLL
transitions. i.e. KLILI. KLjLz. KL1L3. KLsz. KL2L3 and KL3L3.

3. Chemical Shifts

Binding energy is the most important quantity that can be
measured by XPS because of the highly useful information it
provides. in a large number of cases the binding energies for an
element show “chemical shifts” that are characteristic of the
element's chemical state.

in some cases the chemical shifts change regularly as the
oxidation state changes. For example. aluminum shows chemical
shifts which proceed from low to high binding energies as the
chemical state changes from A10 to A132 The reason that binding
energies shift in this manner is because the ejected electrons from
the A13; are subject to less electron shielding from the valence
electrons and a greater net positive charge from the atom than the
ejected electrons from A10. Unfortunately. binding energy chemical
shifts depend on other factors such as electron relaxation and extra-
and intra-atomic forces. so that one cannot use them to obtain an
unambiguous measure of the oxidation state.

The best way to handle chemical state determination is the
Auger parameter method. This method distinguishes chemical states

by observing the energy separation between the photoelectron lines

11

and the Auger lines of an element. The Auger parameter. a . is

defined by Wagner(33) as.

a - KE(A) - KE(P)
in which KE(A) is the kinetic energy of the Auger electron.
KE(P) is the kinetic energy of the photoelectron.
For insulators. on is independent of static charging and is
characteristic of a particular chemical state. The disadvantage of this

definition is that it can result in negative values. A modified Auger

parameter. a'. is then defined as

a' - on + 170- [(801) + BE(P)
in which BE(P) is the binding energy of the photoelectron.

The modified Auger parameter is not only independent of
charging. but also is independent of the energy of the X-ray source.
This eliminates the need for an outside binding energy reference
because the chemical states are determined by the peak separation.
rather than by the absolute binding energies.

The reason that a reference such as the C15 line is not needed
when Auger parameters are used is because the chemical states are
not determined by referencing to a standard binding energy.
instead. a two dimensional array can be constructed as shown in
Figure 1-4 by plotting the binding energy of the most intense
photoelectron line on the x-axis and the kinetic energy of the most

intense Auger line on the y-axis for a series of known compounds.

12

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

m 1 1 I TI 1 1 Uh T j V T—r T I I I TV I r T—ftj‘ ”1‘
b q
1' d
" -1
994 2013
p c-
"' -1
P «1
" -1
993 2012
h .1
I- v ‘
- J
992 2011
r- ..
1' -1
b d
- é? -
991 2010
5 c -
g It
5 " 1
- Ci
“‘ 99° *2000
Q - ’ q
'- c1
1— d
2’ 909 V m
a P
0 " -1
* d
3' at ,«r 1 .
980
v“ .
d
d
907
900
”5 l
mmnumuwtta ‘1
mews-ouch..." "‘
in. ‘
1 1 l l l L l l l l j l I l l 1 l 1 LL 1 1*

 

10a 1025 1024 1023 . 1M2 1021 1“ 1019
29* BlNDlNG ENERGY. 0V

Figure 1-4 Two-dimensional zinc chemical state plot for XPS
(from tabulation by C. D. Wagner. Handbook of X-ray
Photoelectron Spectroscopy: Perkin-Elmer: 1979: p84)

AUGER PARAMETER PLUS PHOTON ENERGY

1.3

Characterization is performed by location of a point on the two-
dimensional array instead of determining the absolute binding
energies from the photoelectron lines and Auger lines. Errors in
charge referencing introduce uncertainty in data points parallel to
the Auger parameter grid. For this reason. labels of compounds are
shown as rectangles with the long dimension parallel to the grid
lines. Since the function plotted is (01 + [10). rather than (01) itself. no
notation of the photon energy is required and the information can be
utilized with BSCA data obtained by using any X-ray source.

For example. the binding energies of the strongest
photoelectron and Auger electron peaks of Zn and mo are 1021.4 eV
(Zn 2p3/z). 261.3 eV (Zn L3M45M45). 1021.6 eV (ZnO 2p3/2). and
264.8 eV (ZnO L3M45M45) using Mg K01 X-rays. There is only a 0.2
eV shift between the Zn 2p3/2 photoelectron lines of Zn and ZnO. but
the Auger parameter of zinc metal is 760.1 eV (Zn 2p3/2 - Zn
L3M45M45) while that of ZnO is 756.8 eV (ZnO 2133/2 - ZnO L3M45M45).
Therefore. the difference between the Auger parameters of Zn and
ZnO is 3.3 eV. This is much better than the 0.2 eV shift between the
photoelectron peaks of Zn and ZnO.

However. the Auger parameter method is not applicable to all
elements. in some cases. natural peak widths of Auger lines are
larger than photoelectron lines. resulting in less accurate
measurements in line shifts. At present. sufficient data for the two
dimensional chemical state plots are available only for nine

elements.(F. Na. Cu. Zn. As. Ag. Cd. in. and Te)

. ”5.-.!

 

14

(ll). Powder X-ray Diffraction Line Broadening

interatomic spacings in crystals are on the order of l A. The
wavelength of X-rays is of the same order. Hence. crystals can act as
diffraction gratings for X-rays. This was first realized by von Laue in
1912 and forms the basis for the determination of crystal structures.
The Bragg equation is a fundamental equation of powder X-ray
diffraction. and is given by

2dsin0 - 11).. 11-1. 2. 3....

Constructive interference between waves scattered by the lattice
points produces a diffracted beam of X-rays only for the angles of
incidence that satisfy this equation. With the known wavelength 2. of
X-rays. the interatomic spacing d can be determined by measuring 0.

This is illustrated in Figure 1—5.

x-ray

 

e——e—-e—e——e

Figure 1-5. Diagram of X-ray diffraction.

1f the particle sizes in the sample are less than 1000 A. the
peak width tends to be broadened and the smaller the size. the
broader the peak. The average crystallite size of a powder sample
can be estimated from the broadening of the powder X-ray line by

using the well-known Scherrer equation

15

K A
Boose

Lhkl"

 

in which L is the average particle size along the direction of the
Miller indices(hkl).
i. is the wavelength of the X—rays used.(eg. Cu K01 is
1.540 A)
K is Scherrer's constant and has a value of about 0.9.
0 is the Bragg angle.
[3 is the peak width at half height in radians.
Bragg has given a simplified derivation of the Scherrer equation that

employs only the ordinary principles of optical diffraction.(34)

(Ill). Electron Diffraction

When a beam of electrons passes through a sample in the
electron microscope. some of the electrons are deflected or scattered
from the main beam in the same way as X-rays. However. the Bragg
angle is much smaller. because the wavelength of the electrons is
ordinarily much smaller than that of X—rays. For example. under
normal operating conditions the wavelength of the electrons is
0.00370 nm when the accelerating voltage of a transmission electron
microsc0pe is 100 kV.

The diffraction pattern from a crystalline specimen is a record
of the periodic structure or repeating array of that specimen. If the

specimen is a single crystal (that is a single repeating array of atoms)

16

then the diffraction pattern will consist of an array of bright spots.
If the specimen consists of a large number of small discrete areas.
each with exactly the same atomic array. but at different orientations
to each other. the specimen is termed polycrystalline. and its
diffraction pattern will consist of a series of concentric rings.

We now consider the electron microscope as a simple electron
diffraction camera. with the electron beam striking a specimen and
being diffracted to form a diffraction spot on the photographic plate
at distance R from the center of the diffraction pattern. as shown in

Figure 1-6.

incident electron beam

  

. —>
spemmen

camera length L

photographic plane

 

Figure 1-6. Diagram of electron diffraction.

The distance between the specimen and the plate. called the
camera length. is designated by L. and by simple geometry
tan20 - R/L

The Bragg Law states that
2dsin6 - 71.

 

l7

and since the angles 0. through which the electrons are diffracted are
very small. (only 1'—2') the approximation

tan20 s.- sin20 a Zsine
can be made with very little error. This gives

R/L - k/d
or

Rd - )(L.
Thus. if values of R. L and 2. for a particular diffraction spot or ring
can be measured. the d-spacing of the lattice planes that give rise to

that spot or ring can be determined.
D. Objective

The objective of this study was to test the generality of
homogeneous reductions of soluble metal salts by solvated electrons
and/or alkali-metal anions in aprotic solvents. Although many of the
metals studied can be prepared as small particles by other methods.
this method has the advantage of being rapid and quantitative and of
requiring only low temperatures for all steps. The method has
potential applicability to metal and alloy formation in the form of
powders. on inert supports. and in the pores of zeolites.

Rapid reduction to the zero oxidation state produces metal
particles that form colloids and aggregates with crystallite sizes that
range from less than 3 nm to about 15 run. When compounds of two
different metals are used. intermetallic compounds or alloys form
rather than simple mixtures of the separate metals. The initial

reduction products must consist of very small. reactive particles. as

 

18

binary alloy formation is observed rather than simple mixtures of
metals when two metal salts are simultaneously reduced. in this
thesis. 1 report the formation and identification of such compounds
and alloys. and the reductive preparation of some metals in zeolite

pores.

Chapter 2 EXPERIMENTAL METHODS
A:Synthesis

The reduction took place in an H-cell with a medium frit as
shown in Figure 2-1. The cell was attached to a vacuum line until
the inner pressure was less than 2x10‘5 torr. The H-cell was then
loaded into a helium-filled dry box. A small sample (1-10 mg) of the
desired compound was added to the H-cell in the dry box. After
removal from the dry box. the H-cell and an ampoule of pre-
synthesized alkalide or electride (about 0.5 mlllimole) which had
been kept cold in liquid nitrogen. were transferred to a glove bag
contained a nitrogen atmosphere. The ampoule of alkalide or
electride was cut off and the contents were poured into the other
side of the H—cell. A liquid nitrogen bath was used to cool the cell to
prevent decomposition of the alkalide or electride.

The H-cell was once again placed on the vacuum line. It was
kept cold by immersion in a bath of liquid nitrogen until the cell had
been pumped down to about 10'4 torr. and then a bath of
isopropanol to which dry ice had been added to keep the
temperature at about -50 'C was used to keep the cell cool. About 20
mL of prepurified dimethyl ether or tetrahydrofuran was then

distilled into each side.

19

 

20

 

 

 

 

 

 

 

 

 

 

 

   

Figure 2-1.Kontes double tube('H-—cell") use to prepare metal and

alloy particles

21

The solvent had been stored under its vapor pressure after
pumping out the nitrogen gas to about 10‘5 torr. (freeze—pump-thaw)
Enough blue solution of the alkalide or electride was poured through
the frit to react with the solution of the metal salt to be reduced after
both solids had been completely dissolved in the solvent.

The reaction was complete immediately after the addition of
alkalide or electride as indicated by fading of the blue color. A slight
excess of alkalide or electride was added until the blue color no
longer disappeared to make sure that the reaction was complete.
Different colors of various colloids were formed at the same time.
The solvent was then distilled off under vacuum. The products were
removed from the walls of the H-cell and mounted on indium or lead
foil in the He-filled dry box. A vacuum transfer vessel. which was
designed to transfer chemically reactive specimens from a controlled
atmosphere to an analytical system. was used to carry the sample
from the dry box to the XPS transfer chamber. The pressure of the
main chamber of the XPS instrument was kept below 3x10‘8 torr
during measurement.

in some cases deionized and degassed distilled water or
prepurified methanol was used to wash away the by-products in a
glove bag. The products were removed from the cell and were
placed into a test tube. About 10 mL of washing solvents was then
added into the test tube inside the nitrogen-filled glove bag.
Washing was performed by centrifugation to separate the
undissolved metallic particles from the water- or methanol- soluble
by-products. XRD patterns were recorded 'by placing the the

precipitate on a glass slide either with or without washing. A drop of

22

washed suspension was put on the TEM carbon-coated grid and
allowed to dry in the air. after which TEM micrograghs. EDS and
electron diffraction patterns were made.

The aprotic solvents MezO or THF were distilled from solutions
of excess Na-K and benzophenone into stainless-steel (MezO) or glass
(THF) storage vessels. Liquid 1505 crown ether was purified by
distillation. All the alkali metals and metal salts were purchased

from Aldrich or AESAR in the highest available purity without any
further purification. Liquid samples were TiCl4(99.995+7.).

GeCl4(99.9997.). SiCl4(99.999%). SbC15(99%) and SnCl4(99.999%).
Solid samples were CuC12(99.99997.). GaC13(99.99+%). Zn12(99.99%).
MoC15(99.99%). FeCl3(98%). HthCl6(99.9%). inCl3(99.999%).
TeBr4(99%). WCl6(99.9+%). AlC13(99.99%). VC13 and AuC13. The oxide

support was neutral activated A1203 with a specific surface area of

155 mZ/g and 150 mesh size. Either a 0.25 mm thick 99.99+%
indium foil or a 0.5 mm thick 99.99% lead foil was used to mount the

samples for XPS studies.

B. X-ray Photoelectron Spectroscopy

X-ray Photoelectron Spectroscopy (XPS) or Electron
Spectroscopy for Chemical Analysis (ESCA) was performed with a
Perkin-Elmer PH] 5400 ESCA/XPS spectrometer system. its analyzer
is a Spherical Capacitor Electron Energy Analyzer (SCA) with an
Omni-Focus ll lens and a small area lens (200 um dia to 3 mm x 10
mm with 4 selectable apertures). This permits efficient small area

analysis without sacrificing overall system analytical flexibility for

23

Auger parameter and photon dependent depth profiling studies.
Unlike a focused X-ray source. the small spot ESCA design gives an
energy resolution that is independent of the analysis area.

The most commonly used anode material for ESCA/XPS is
magnesium which has a K01 X-ray energy of 1253.6 eV. it has a
narrow natural X-ray line width which facilitates chemical species
interpretation and is a highly efficient source of X-rays at moderate
power. Photoelectron lines generated by Mg K01 X-rays are narrower
and more intense than the same lines generated by any of the other
commonly used non-monochromated X-ray sources.

The system is also equipped with a monochromator to deliver
the highest available energy resolution. in which case. Al K01 (1486.6
eV) X-rays are used. The monochromator provides maximum X-ray
collection efficiency for unsurpassed counting rates. This
monochromator X-ray source works with the SCA and Omni-Focus ll
lens to also provide high signal-to-background ratios for superior
trace element detection. sharp. narrow line widths for enhanced
chemical speciation. and the elimination of unwanted satellite lines

for unambiguous data interpretation. All data were recorded by

scans at 15 KV. The power of the X-ray sources was 600 W for Al K01
and 400 W for Mg K01.

A vacuum transfer vessel was used to transfer chemically
reactive specimens from a controlled atmosphere to the analytical
system. or from one system to another without atmospheric
exposure. The specimens were transported under an inert gas
atmosphere or under vacuum. This vacuum transfer vessel allows

one to transport a sample of all products of the reduction from the

 

24

helium filled dry box to the introduction system of the XPS without
oxidation.

There is also a differential ion gun on the XPS system to
generate an energetic ion beam for sputter etching of solid surfaces.
During operation. an inert gas such as argon is bled into the gun
ionization chamber. A heated tungsten filament emits electrons
which are attracted to a positively biased grid and accelerated
through the gas in the chamber.

ions are created by electron impact within the ionization
chamber. extracted from the chamber toward the focusing and
deflection lenses. and directed onto the specimen. When the surface
is oxidized. this argon ion sputtering technique is used to probe the

sub—surface composition.
C. Powder X-ray Diffraction

Powder X-ray diffraction (XRD) provides direct evidence for
metal particle formation when the crystallite size is greater than 3
nm. The width of the strongest XRD line can be used with Scherrer's
equation to determine the average size for crystallites between 3 and O
20 nm.

For X-ray amorphous metals. common in this work. XRD only
provides an upper size limit of about 3 nm. but cannot be used to
verify metal formation. The by-product salts were always
crystalline. so that XRD provides a convenient measure of the
efficiency of removal of these salts. By contrast. organic

decomposition products are not usually detectable by XRD.

25

A Rigaku D/max—RBX rotating-anode diffractometer equipped
with a scintillation counter detector and a graphite monochromator
to yield Cu K01 (wavelength 1.54184 A) radiation was used under the
control of a DEC Microvax computer. All data were recorded by scans
at 45 KV and 80 mA.

D. Transmission Electron Microscopy

Transmission Electron Microscopy (TEM) was used to obtain
particle sizes and the morphology of aggregation. A JEOL 100 CX II
transmission electron microscope operating at 100 KV was used for
imaging. energy dispersive spectra (EDS) and selective area electron
diffraction (SAD).

It was not p0$sible to prevent air oxidation with the
instrument used. so that reactive metals could not be directly
observed in the TEM. But the particle size distribution of the
products of oxidation could be studied. it was also possible with this
instrument to obtain elemental analyses of microscopic crystallites
by EDS. This allowed us to verify the presence of both components of
an alloy and to identify regions that contain only one metallic
element. For products that could be briefly exposed to air. SAD

provided identification of the structure.

 

Chapter 3 SINGLE ELEMENT REDUCTION
A. Reduction with Alkalides or Electrides

Alkalides and electrides owe their parentage to alkali metal
solutions in ammonia. amines and polyethers.(35) The species e‘solv
and M" have been extensively studied in solution.(36.37.38) Indeed.
reductions that involve such species(“dissolving metal reductions”)
have been used in the laboratory and industrially for many years.
The unique feature of alkalides and electrides is the ability to
produce stoichiometric crystalline salts in which trapped electrons or
alkali metal anions serve as the anions. This Is achieved by using
powerful cation complexants such as crown ethers or cryptands to
protect the alkali cation from reduction by e'sojv or M'. To first
order. most solid alkalides and electrides may be viewed as close-
packed large (8-10A diameter) cations with e'solv or M" in the
holes that are produced by such packing.

The organic complexant not only permits isolation of solvent-
free crystalline alkalides and electrides. but also serves to greatly
enhance the solubility of alkali metals in such aprotic solvents as

Mezo and THF. The relevant equilibria are:

26

 

27

 

 

M(S) ——r M3011, + 8 solv (1)
2 M (s) ——r M 301v + M solv (2)
Mloiv + nL 4————7 WM: (3)

in which L is the complexant and n is l or 2. Reaction 3 lies far
to the right. which tends to drive reactions 1 and 2 to the right also.
The effect on metal solubility is dramatic. For example. potassium is
not perceptibly soluble in Me20 (10'5 M solutions would be easily
discernable by eye because of the intense colors of e’solv or M'). In
the presence of stoichiometric amounts of 15C5. at least 0.5 M
solutions of K*(15C5)2K' are formed at -25 °C. The solubility is
presumably even higher than this since no precipitate forms upon
cooling this solution to -78 'C. This dissolution process is slow.
however. and potassium powder or films require about one hour at
-25 'C to form concentrated solutions. By contrast. pre-synthesized
alkalides and electrides dissolve rapidly.

The solvated electron (e‘solv) is thermodynamically the most

powerful reducing agent possible in a given solvent. it also usually
reacts rapidly with metal ions and with simple compounds that
contain a metal in a positive oxidation state. Alkali metal anions.
M" . are nearly as effective and can provide two electrons in a single
encounter. Side reactions with the solvent arise. however. when
protic solvents such as ammonia or a primary or secondary amine is
usedI39)

A solution that contains e'sojv or M‘ in an aprotic solvent is an

attractive reducing medium. The alkali metals are solubilized in

28

aprotic solvents such as dimethyl ether or tetrahydrofuran by using
a suitable cation complexant such as a crown ether or a
cryptand.(40.4l.42.43) The resulting homogeneous solutions contain
either ego“, or M" depending on the metal/complexant ratio.(44)
Identical solutions are prepared by dissolving pre-synthesized
electrides or alkalides in the solvent. When a metal salt or complex
are dissolved in the same solvent. rapid reduction to the metal can

occur upon mixing the two solutions.
B. Reducing agents

Six alkalides and two electride were used as reducing agents.
These include: K+(18C6)Na‘. Cs+(18C6)2Na'. Cs+(15C5)2Na'.

Rb+(15C5)2Na‘. K*(15C5)2K“. Rb+(15C5)2Rb'- K+(15C5)2e'. and

Cs*(15C5)2e‘. It is not important which alkalide or electride is used

since it simply acts as an electron donor. There is no perceptible
difference in the reducing ability among them as the blue colors of
the alkalide or electride solution disappear immediately when the
reaction occurs. The compounds K*(15C5)2K‘ and K*(15C5)2e‘ were
most often used as reductants because of their high solubilities and

ease of preparation.
C. Results
This work presents a new method for preparation of small

metal/oxidized metal particles that utilizes homogeneous reduction

by dissolved alkalides or electrides in an aprotic solvent such as

29

dimethyl ether or tetrahydrofuran.(45) Salts of Au. Pt. Cu. and Te
form metallic particles with little or no oxidation even when washed
with methanol. Reduction of salts of Ni. Fe. Zn. Ga. 81. Mo. W. In. Sn.
Cd. and Sb yields surface oxidation over a a metallic core. while only
oxidized Ti was observed.

The procedure for the preparation of small metal particles was
described in Chapter 2. While nearly any soluble metallic compound
is reduced by alkalides or electrides in Mezo or THF. the less reactive
noble metals. Au. Pt. etc.. are easiest to isolate and characterize. Gold

was used to test the method and to develop the methodology because

 

AuCl3 is highly soluble in MezO and can be easily obtained as the
anhydrous compound. and because the product. Au metal. does not
oxidize easily when exposed to the air while almost all other metals

do when they are in the nanophase state.

(1) Gold
AuCl3 was reduced by K*(1505)2K‘ or K*(15C5)2e‘ in MezO or
THF. The reaction stoichiometries such as
2 AuClB + 3 K+(15C5)2K‘ fle‘ 2 Au + 3K"(15C5)2C1' + 3KCl
and '

M o
AuCIB + 3 K+(15C5)2e‘ #— Au + 3 K”(15C5)2C1‘

were verified by XRD studies of reaction products after solvent
removal by distillation. The by-products. K*(15C5)2C1‘ and/or KCI.
were washed away by methanol or water. The remaining black

products are pure small metallic gold particles confirmed by XPS.

30
XRD. SAD. and EDS. AuCl3 was reduced 11 times and the results were
reproducible.

For crystalline particles of average diameter larger than 3 nm.
the mean diameter was determined byapplying Scherrer‘s equation
to the XRD line widths. TEM micrographs were also used to
determine particle size and size distribution. EDS permitted analysis.
and SAD was used to identify the bulk structure of the particles. XPS
(photoelectron and Auger peaks) identified the surface composition
(metallic or oxidized). and when necessary. argon ion sputtering
removed the surface layer to verify the presence of subsurface
metal.

Crystallite sizes (and particle sizes) were of the order of 10 nm.
so that XRD identification was possible. When dilute solutions (10'2
to 10'3 M) of AuCl3 were reduced. colloidal gold that only slowly
precipitated was formed in MeZO. The particle size was smaller
when more dilute solutions were used. After removal of MezO by
distillation. the residue again formed a colloidal solution when taken
up in water.

Figure 3-1 shows the XPS spectrum of Au (4f7/2 and 4f5/2)
from the product of reduction of AuCi3. The binding energies of both
the Au 4f7/2 state at 83.70 eV and the 413/2 state at 87.30 eV show
that the product is pure metallic gold.(46)

Figure 3-2 shows typical XRD lines that were used to determine
average particle sizes for Au. The average particle size estimated
from the line width and Sherrer's equation is 10 nm. Figure 3-3
shows the size distributions for Au from an electron micrograph

(Figure 3-4). There is very good agreement between the average

31

crystallite size from XRD and the particle size distribution from EM.
The SAD patterns shown in Figure 3-5 were also taken on the same
area. All the rings in Figure 3-5 were converted to d-spacings and

were compared to a standard XRD Au file as shown in Table 3-1.

 

 

L A A

I L A L
‘ T T r T 1'

L A L
I l I V
i

 

 

 

['11 u.-
e—rvu; owoao

 

 

Binding Energy. eV

Figure 3-1. XPS spectrum of Au from the product of reduction
of AuCl3.

32

1000 '
i
ll I
‘1 , F‘
a’JG - I

 

o +w-r‘r-‘r-ihrr r-w-f Y'v-r~v-I'1-rw-1'-{ s Y’P'r—F"V’T'r—T‘+rr1 ..., +"r v .. v-f v 1 e 1 E - v -
37.0 37.5 38.C 33.? 39..

Figure 3-2. XRD lines(l l 1) used to determine average particle

sizes for Au.

40'

30'

20

10'

33

 

 

 

24 48 72 96 120144168>l92

Particle size. A

Figure 3-3. Particle size distributions for Au.

34

 

Figure 3-4. Electron micrograph of Au.(mag-100.000X. 1.000A/cm)

 

Figure 3-5. Selected area electron diffraction of Au.

(camera length-83 cm)

L.

-—I\

 

 

Table 3-1. Electron diffraction of Au from reduction of AuCl3,

35

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

D(cm)a d(A)b d/d'c 'd'(A)d 11x16
1 2.74 2.24 0.952 2.355 111
2 3.14 1.96 0.959 2.039 200
3 4.51 1.36 0.944 1.442 220
4 5.28 1.16 0.943 1.230 311
5 5.57 1.10 0.934 1.1774 222
6 6.44 0.954 0.935 1.0196 400
7 6.98 0.880 0.940 0.9358 331
8 7.18 0.855 0.938 0.9120 420
9 7.84 0.783 0.941 0.8325 422

 

 

 

 

a. Ring diameter of electron diffraction.(camera length - 83 cm)
b. d-spacing converted from ring diameter.

c. d-ratio from electron and powder X-ray diffraction.

d. d-spacing from powder X-ray diffraction files.(4-0784)

e. Miller index.

36

This method was also used to prepare highly dispersed metals
on oxide surfaces as previously done by other methods.(47) To
verify this. AuClg, in Mezo was adsorbed on neutral activated A1203
(155 mZ/g) and reduced with K*(15C5)2K' in Mezo. Au particles
(about 6 run average diameter) were randomly dispersed on the
surface as shown in Figure 3-6. One may be able to prepare a large
surface area metal oxide catalyst by co-reduction of a noble metal
and an active metal. followed by oxidation of the latter. it should be
also possible to prepare organometallic compounds by reaction of

suitable precursors with freshly prepared active metals.

 

Figure 3-6. Electron micrograph of Au particles on A1203.
(mag-190.000X. 550A/cm)

37

A gold colloidal solution was formed by adding water to the
reaction products after distilling out the MezO. The average diameter
of Au produced in this way was about 10 nm as determined from
XRD. The optical absorption spectra of a gold colloid solution (ruby
red) is shown in Figure 3-7. The 520 nm peak was assigned to
surface plasmon absorption by Abe et al.(48) This colloid is very
stable over a period of several months. This optical peak was also

found in deposited particles (12 nm diameter) in glass.(49)

 

About..-

 

 

 

Figure 3-7. Optical absorption spectra ofanAu colloidal solution.

38

In an attempt to control the particle size. two AuCl3 solutions
with different concentrations were reduced with K*(15C5)2K‘ in
Mezo and the products were collected by distilling out solvent. The
XRD data showed that the higher concentration formed larger
particles of Au. A 1.6x10‘4 M AuCl3 solution formed 5.8 nm Au
particles while the higher concentration (6.5x10‘4 M) formed larger
Au particles (10.5 nm) as shown in Figure 3-8. However. the by-
product crystallite sizes remained almost the same even when

different Au particle sizes formed. This was verified by the same
peak height and width from KCl (111) at 20 - 40.5'.

1900 V !\

1600 ‘P j

300 '
1 I B \\

1000 "

MM

400 ‘
35.0 36.5 37.0 37.5 38.0 38.5 39.0 39.5 40.0 40.5 41.0
2 THETA

ONM\UI-GZO
‘V:

 

 

Figure 3-8. XRD spectrum of the products from the reduction of
AuCl3. (A-6.5x10-4 M. B=1.6x10-4 M)

39

By using all of the characterization techniques described above.
it was often possible to identify the major phase or phases resulting
from reduction and to obtain particle size information. But the high
reactivity of nanoscale metal particles and the presence of organic
complexants and solvents often resulted in the inclusion of organic
decomposition products.

FT-IR spectroscopy was used to demonstrate their presence
and to identify major iR-active groups on the surface of the metal
particles as shown in Figure 3-9. The following peak positions were
identified: 611(C-H or O-H). 1089 and 1117(C-O). 1385(major
impurity from KBr). 1634(O-H). 2852 and 2925(C-H). and 3448(O—H).
A sample of washed Au particles was also sent to Galbraith
Laboratories. Inc. for chemical analysis and it was found that carbon
was 0.70 7. and hydrogen was < 0.5 71 in atomic percent

concentration.

 

40

. fiflNSi‘l I TTfiNCE
7 T ‘l 9 60 71 82

38

27

 

 

211 00 3590 3180 2770 2360 13—50 1540 1T30 1520 310
unveuuneen

Figure 3-9. FT-IR spectra of washed Au particles.

 

41

(11) Copper

Although bulk copper is very stable in the air. it is easily
oxidized at room temperature when the particles are of nanometer
size. CuClz was reduced in MezO by the present method to produce
small metallic copper particles. These were fully characterized by x-
ray powder diffraction. electron diffraction. and x-ray photoelectron
spectroscopy. Only metallic copper peaks were detected by XRD from I"
the precipitates after washing away the by-products. The mean

particle size of the strongest peak of Cu(lll) from the XRD line

_\ 1|~_rl-'r":l-

broadening. as shown in Figure 3-10. was about 57 A after washing

with methanol.

Q

. :1,
1 #11“va
4, (I

 

L} 1 ()1
W W MMM .11 r
. 111W

 

 

41.5 42.0 42.5 43.0 43.5 44.0 44.5 45.0 45.5

20
Figure 3- 10. XRD lines(111) used to determine average particle

sizes for Cu.

42

Figure 3—11 shows the characteristic electron diffraction
pattern of metallic copper. with lines (111). (200). (220). (311) giving
rise to rings from inner to outside rings. All the rings in Figure 3-11
were also converted to d-spacings and were compared to a standard
XRD Cu file as shown in Table 3-2.

 

Figure 3-1 1. Selected area electron diffraction of Cu.

(camera length - 140cm)

 

43
Tabie 3-2. Electron diffraction of Cu from reduction of CuClz

 

D(cm)a 0(A1b d/d'C d'(A)d hkle
1 4.93 2.10 1.0057 2.088 111
5.66 1.83 1.0120 1.808 200
8.02 1.29 1.0094 1.278 220
9.38 1.10 1.0092 1.090 311
9.81 1.06 1.0153 1.044 222

 

 

 

 

 

 

 

 

 

 

 

 

MAO-IN

 

a. Ring diameter of electron diffraction.(camera length - 140cm)
b. d-spacing converted from ring diameter.

c. d-ratio from electron and powder X-ray diffraction.

d. d-spacing from powder X-ray diffraction files.(4-0836)

e. Miller index.

The electron micrograph of Figure 3-12 agrees well with the
average particle size of 57 A obtained from the XRD line broadening
of the metallic Cu product from the same run. Figure 3-13 shows the
particle size distribution of these metallic copper particles. These
particles were also characterized by energy dispersive spectroscopy
as shown in Figure 3-14. Only Cu peaks appear on the Ni grid.

The small Cu particles were also characterized by XPS and only
the 2p3/2 peak at 932.6 eV and the 2p1/2 peak at 952.4 eV appear
in Figure 3-15-A (before washing) and Figure 3-15-C (after washing
away by-products with methanol). Figure 3-15-B shows that a
small amount of CuO is present on the Cu surface after exposing the

sample to air for 24 hours.-

 

44

 

Figure 3-12. Electron micrograph of Cu.
(mag-320.000X. 310A/cm)

30'

20'

 

 

 

12 25 37 50 62 75 87

Particle size. A

Figure 3-13. Particle size distribution of Cu.

>100

45

 

—.-u-—

 

 

8 f
8 3

 

46

o M _ o M
221122121121 1 29212129212

11 . r 2 12.1.1.

 

 

 

 

(B)
N(E)/E

 

 

(c) /\

N(E)/E WW \

 

 

 

 

 

 

 

L L A A A L
A v
'7 y r v

$3.0 $4.. ”.0 $1.. 90.. ”1.. 935.0 ”4.. 325.0

Binding Energy , eV

Figure 3-15. XPS spectrum of Cu (A before washing. B exposed

to air for 24 hours. and C after washing)

47

(Ill) Tellurium

TeBr4 was reduced by K*(15C5)2e' in dimethyl ether. Figure
3-16-A shows the XPS results of the 3d5/2 and 3d3/2 peaks of

metallic Te before washing. After oxidizing these products in air for
3 min. the oxide peaks were larger than the metal peaks as shown in
Figure 3-16-B. These particles were partially oxidized after washing
with methanol as shown in Figure 3—16-C. Clean metallic Te peaks
were observed in another methanol—washed product.

Only metallic Te peaks were observed by powder X-ray
diffraction. Although the average particle size estimated from line
broadening in Figure 3—17 for Te (101) is 117 A (line width at half
height [3 - 0.7). the micrograph (Figure 3-18) shows interesting rod-
like shapes . The electron diffraction pattern (Figure 3—19) shows
the hexagonal structure of metallic Te with a-4.46 A and c-5.93 A.
All the rings in Figure 3-19 were converted to d-spaclngs and were

compared to a standard XRD Te file as shown in Table 3-3.

48

 

 

 

 

0 M 0 M

3d8/2 31.342 3d5/2 3d5/2
(A) .
NiE)/E

«W‘J

,N

(B) /
N(E)/E

 

 

(C) . /
N(E)/E AWL/

 

 

 

 

 

 

 

 

A A A A A A A A
A
V V V V Y V V ' ‘ : ¢
Y
S‘J 91.. ”.0 ”.0 573.. 9&0 91.. S72 0 S
0

Binding Energy , eV

Figure3-16. XPS spectrum of Te (A before washing. B exposed

to air for 3 min. and C after washing)

49

300

--. +... .__- ‘

._-+. —-.

240

180

120

. .__... -__.+._.--_+.._~. .. +-..... .-

60

... .+.._

O
f
I

l

0 er-s ri-v1rrf-r-t-r'ri-m-rfl-vv-ri-rtw-r-frfTri‘t-Tfrhflfi fi-v-r-rrf‘r-c-rvhmfv-t 1 1+“
26.0 26.5 27.0 27.5 23.0 28.5 29.0.

20

Figure 3-17. XRD lines(100) used to determine average particle sizes

for Te.

 

50

 

Figure 3-18. Electron micrograph of Te.
(mag-140.000X. 700A/cm)

 

Figure 3-19. Selected area electron diffraction of Te.

(camera length = 140cm)

Table 3-3. Electron diffraction of Te from reduction of TeBr4,

51

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

D(cm)a 0(A1b d/d'c d'(A)d hkle
1 2.74 3.781 0.979 3.86 100
2 3.29 3.149 0.975 3.23 101
3 4.53 2.287 0.973 2.35 102
4 4.76 2.176 0.976 2.23 110
5 5.12 2.023 0.968 2.09 111
6 5.41 1.915 0.967 1.98 003
7 5.79 1.789 0.975 ' 1.83 201
8 6.06 1.710 0.973 1.76 103
9 6.59 1.572 0.973 1.616 202
10 7.24 1.431 0.967 1.479 113

 

 

 

a. Ring diameter of electron diffraction.(camera iength- 140cm)
b. d-spacing converted from ring diameter.

c. d-ratlo from electron and powder X-ray diffraction.

d. d-spacing from powder X-ray diffraction files.(4—0554)

e. Miller index.

52

(1V). Nickel
By adding triethylphosphine to increase the solubility. NiBrz

was reduced by K*(15CS)2K‘ in dimethyl ether. The products were

kept in the H-celi under vacuum for four days and analyzed by XPS.
Figure 3-20-A shows only metallic Ni peaks. Small NiO peaks appear
when these particles were oxidized in air for 24 h. but all Ni particles
were oxidized after washing by methanol as shown in Figure 3-20-B
and 3-20-C.

(V). Antimony
SbC15 was reduced by K+(15C5)2K' in dimethyl ether. Both

metallic and oxide 3d5/2 and 3d3/2 XPS peaks appear in Figure 3-

Zl-A and 3-21-B before and after washing by methanol. The
washed products were sputtered by Ar ions for 20 min. The result
shown in Figure 3-21-C indicates that only a small amount of oxide

was left after sputtering.

(VI) Gallium

GaCl3 was reduced by K*(15C5)2K‘ in dimethyl ether. Both
metallic and oxide L3M45M45 Auger peaks appear in Figure 3-22-A.
The sample was bombarded by argon ions for 2 min and 5 min as
shown in Figure 3-22-8 and 3-22-C respectedly. it is clear that the
longer the argon ion sputtering the higher the intensity of the

metallic peak and the lower the intensity of the oxide peaks.

 

53

O M 0 M
2p 1/2 2p 1/2 2p3/2 2p3/2

 

(A) .
N(E)/E ’ ' ‘ ,

 

 

 

 

 

 

 

 

 

 

 

 

 

L L
v

A L
fi‘f

A A J A A A A A
Y Y V V V v V Y i Y
000.0 004.0 070.0 011.0 “3.0 “4.0 60.0 61.0 ”.0 0

A
i T

._ Binding Energy , eV

Figure 3-20. XPS spectrum of Ni.(A) before washing (B) oxidised in
the air for 24 h (C) washed by MeOH

t _ .

54

O M O

3d3/2 3d3/2

4
v

M

311512 3d5/2

 

(A)
N'(E)/E

\

\

A
f

 

A _L- A .l A . A
Y

 

(B)
N(E)/E

 

 

 

 

 

O
4 #%A
V

 

 

 

A
v V

J
Y

L A A
Y I V V Y V
531.0 53.0 533.0 53L0

Binding Energy , eV

A A
V V
O
90.0

52]) 525.

Figure 3-21. XPS specrum of Sb. (A) before washing (B) washed by

MeOH (C) sputtered by Ar ions

 

 

55

o u
LSHQSHU 1.311,. 3‘“:

 

 

 

 

 

 

 

 

 

. A : : c 4. 4. ¢ : Af—t %% + 4. t . a L L
i 1
, 0
o 1b
Ii 2
(A) '1
.‘ l
5+
+
N(EHE 4M
3 .1) 5% hA/‘fl
1 ¢ , .
2+ V .2"
”L
a . A - 1 A 2 4 4
I *——¢ ¢ 4 T 4 + 41—4 + ‘ g* 4r T
st 4:
1 11
u 3)
9
(B) 7 4.
. 1.
s 11
4)
"(3),! i .L
5 1*
WM *'
2 j V 1 i’
J»
‘ -
1 A
I . f c t c c : 4. # ¢ ¢ 4 4. : ¢ ‘r : v
10 L L 4 4 + 4. ¢ : 4 ¢ : t ‘r 4. 4 4 t *
5
. 1“
(C) I.
0
Run/3 ‘
3
NM.“
2
l INM~Vva *
T .
i A
I 1 4— f ¢ : —.L t f 4 4 4 L —-+—-—

 

201.5 202. 0 100.5 107.0 151.5 19.0 100.5 107. 0 104. 5 1‘1

Binding Energy . 0v

§

Figure 3-22. XPS specrum of L3M45M45 Auger peaks of Ga. (A)
reaction products (B) sputtered by Ar ions for 2 min (C)

sputtered by Ar ions for 5 min

 

 

56

(V11) Zinc

anz was reduced by K*(15C5)2K‘ in dimethyl ether. Both
metallic and oxide L3M45M45 Auger peaks appear in Figure 3-23-A.
The sample was bombarded by argon ions for 2 min and 4 min as

shown in Figure 3-23-B and 3-23-C respectedly.

(Vlll) Molybdenum
M0C15 was reduced by K*(15C5)2K‘ in dimethyl ether. Both

metallic and oxide 3d5/2 and 3d3/2 XPS peaks appear in Figure 3-

24-A. The sample was bombarded by argon ions for 10 min and 30
min as shown in Figure 3-24-B and 3-24-C respectedly. The oxide on
molybdenum is harder to remove than that on zinc or gallium so that

it took a longer time to sputter off the oxide.

(IX) Tin
SnCi4 was reduced by K*(15C5)2K‘ in dimethyl ether. Both

metallic and oxide 3d5/2 and 3d3/2 XPS peaks appear in Figure 3-

ZS-A. The sample was eXposed to the air for 30 seconds and the
intensity of metallic peaks was decreased to only about half as

shown in Figure 3-25-B.

57

0 .H

 

(A)

rim/1;

 

~u¢un~.o.
A AA‘

 

 

 

A_

LA

A A
V Y I V V V

(3)

Y

«o‘coo'9—
5
f

LLLLAAA
Vv

Rim/1:

 

 

{L 1. .
4L 4

(C)

N(E)/E

   

 

 

 

o—uuaun~.o.ou~u¢

 

 

A A A A A A ' A A
V V '

A A -
v V T V V 1 1
30.0 2.5 90.0 n: no as ‘0 ‘5 no as as

Binding Energy . 0V

_‘ A
Y—v

Figure 3-23. XPS specrum of L3M45M45 Auger peaks of Zn. (A)
reaction products (B) sputtered by Ar ions for 2 min (C)

sputtered by Ar ions for 4 min

 

(A)

A A
v

0

58

0

.M

M.

 

AAA-AAA
VV‘V‘VrTfV

 

“ fl
A A A A
v v v Y

"(M/E

AAA_AAA
7

AA
'VrYVYYY

   

 

V
1
o
1
1

i
y.
)'
V v
p
p
p
p
p

 

A
AA

 

(I)

AAAAA
VVfTYTvvv
44L

tun/n

A A A
T V V V V V V V Y ' V V

A A
v

 

 

 

(C)

uo~ou'.—-Auawnu‘ouuua
r :5?‘

A AA—LLAAALA A
V V V V

rim/1:

“A

 

A L

 

 

 

 

 

   

A
—v’

v

 

L A A A A A A
v r

A A A
v v v V r v v v 1 v

Binding Energy . 0V

A A A
v f

10.0 1&0

Figure 3-24. XPS specrum of 3d peaks of Mo. (A) reaction products
(B) sputtered by Ar ions for 10 min (C) sputtered by Ar

ions for 30 min

59

 

 

 

 

 

 

    

 

o u. 0 M),
343/2 3d3/2 3d5/2 3“5/2
.A r A _ . i
9
O
(A) r
C
S
N<:)/: .
I
2
I
O
I
3
I
(3) '
C
':
N(E)/t 4:
,0
’L
‘1
0 +¢*‘ **‘ ‘ gL;

 

 

 

‘0 Qt. .J .J .J C... “I Q. ‘0

“"‘Itndin; Inert! . 0'

Figure 3-25. XPS specrum of 3d peaks of Sn. (A) reaction products (B)

exposed to the air for 30 seconds

60
D. Summary

The reduction products are indicated in Table 3-4 and may be
classified as follows. .

(1) Washable metals: Metal particles could be washed with
little or no surface oxidation. Compounds reduced include AuCl3.
HthCl6 (hydrated). TeBr4. and CuClz.

(2) Surface-oxidized: Partial or complete surface oxidation
occurs with or without washing. but the subsurface material is
metallic. Compounds include SbC15.ZnIZ. GaC13.SnCl4. lnCl3. and
SiCl4.

(3) Nonwashable: Metal was observed with the unwashed
product only. Oxidation occurs upon washing with methanol.
Included are NiBrz + triethylphosphine. MoClS. FeCl3. and WCl6.

(4) Reoxidized: No metal was observed (with or without
washing). although initial reduction occurs as indicated by fading of
the blue color of the alkalide solution. This includes only TlCl4 so far.
although TaC15 is also readily oxidized by the reaction medium.

Homogeneous reduction of various metal salts with alkalides or
electrides in MeZO or THF at -50 'C or below produces 2-15 nm
diameter particles of metals or oxidized products. Reactions are
rapid. complete. and applicable to a wide range of elements from Ti
to Te. Metallic particles dispersed on high surface area oxides were
also prepared in this way.

The XRD patterns of the unwashed products and the absence of

significant carbon signals in the XPS data of washed samples show

 

61
that most of the crown ether complexant is not destroyed in the
process and could be recovered if desired. This would be important
in the utilization of the method because of the relatively high cost of

the complexants.

 

62

iota—con 3:33 22: .8 :8. 05 com .uonmExoou .v

632—335: .m ”conmvmxoooatam .N ”£82: 03233 J EcuaoEEoE

.23 o. .663» 35820 on. 3.36.? 032 omvotoa 2: «o tam .vé 035.

 

 

 

 

 

 

 

 

 

 

d n ma
2. a an. a. wzs< E .__ 8 3. garbc: 3
fl m N N_N m
oh. am :meU? E 5. 3— 8. 32:2 HN w
m N fl m H v

em 2 commuaNSUMZ 8 omié > C. a

N m m m. m m mm m
m A. mm 2 = _ :5 :> ~> > >_ E
o z o m

 

 

 

Chapter 4 BINARY ELEMENT REDUCTION

A. Introduction

Single element reduction was described in Chapter 3. Soluble
compounds of transition metals and post-transition metals in
dimethyl ether or tetrahydrofuran are rapidly reduced at - 50 'C by
dissolved alkalides or electrides to produce metal particles with
crystallite sizes from < 3 to 15 nm. It is clear that a wide variety of
elements can be prepared as small metallic particles by reducing
dissolved compounds in MezO or THF with M‘ or e‘solv-

Because of the small particle size (and probably even smaller
initial particles) it was of interest to examine the products of
reduction of homogeneous mixtures of two or more metal salts. If
intermetallic compounds could be formed rather than mixtures of
metals. this would open up an opportunity to synthesize new
materials. since many intermetallic compounds that are stable at low
temperatures decompose at high temperatures. A major advantage
of the present method is the ability to control overall stoichiometry

simply by adjusting initial compositions.

B. Experimental

The purified aprotic solvents MezO or THF were distilled from
stainless (MezO) or glass (THF) storage vessels. from which 30-40 mL

63

64

were withdrawn by distillation as needed. The commercially
available H-cell (Figure 2-1) was used to carry out the reduction.
The solid or liquid samples of the compounds to be reduced were
added to one arm and either a pre-synthesized sample of the
alkalide or electride or the alkali metal and a stoichiometric amount
of complexant was added to the other arm in the He-filled dry box.
Small amounts (5-10 mg) of reactants were used to produce
10‘2 to 10'4 M. solutions. Larger amounts up to a millimole were
used when the precursor solubility permitted. The potasside.
K+(15C5)2K‘. and electride. K+(15C5)2e'. were used as the reductants
because of their high solubilities and ease of preparation. To carry
out homogeneous reactions. the desired compound and the alkalide
or electride were separately dissolved in MezO or THF and the
solutions were rapidly mixed. Reduction was immediate to give

colloidal solutions when dilute and precipitates when concentrated.
C. Results

When two different metal salts were used. alloys or
intermetallic compounds formed as indicated by XPS. Confirmation
by electron diffraction was made in the case of air-stable samples.
All binary systems studied so far yielded some evidence for alloy or
compound formation. But all have been X-ray amorphous. thus
preventing identification of the compounds by XRD.

The photoelectron and Auger peaks were used to distinguish
intermetallic compounds from a mixture of the pure metals. but

cannot provide positive identification. When the sample can be

 

65

briefly exposed to air without oxidation. it is possible to use electron

microscopy. EDS and. especially SAD to identify compounds.

(I) Au-Zn

The identification of the known compound AuZn was carried
out by XPS and SAD. Figure 4-1 (A) is the XPS spectrum of Au 4f7/2

and 41'5/2 from the product of reduction of AuCl3. The binding
energy of Au 4f7/2 at 83.70 eV and 4f5/2 at 87.30 eV shows that
the product is pure metallic gold.(50) Figure 4-1(B) is the Au 4f7/2
and 41‘5/2 XPS spectrum from the product of reduction of a mixture

of AuCls and anz that contains excess Zniz. These peaks shift to

84.45 eV and 88.00 eV respectively.

The 4fbinding energy splitting (3.55 eV) from (B) is narrower
than that (3.60 eV) from (A). and shifted to higher binding energy
with respect to the Fermi level. This indicates that an alloy is formed
from the reduction of the mixtures. This result is in good agreement

with the XPS work on Al-Au alloys by Fuggle and coworkers.(51)
Figure 4-1(C) is the spectrum from the mixtures of AuCl3 and anz

with excess AuCl3. This spectrum shows clearly that both pure

metallic gold and alloy exist. The new compound was identified as
AuZn by SAD. Table 4-1 is the comparison of d-spacings between XRD
and SAD.

  

(A)

 

 

 

LAALAAA
v r1 rfrf

 

 

' l
b
‘i
4L
A
l A r

1

 

A L A A
T

‘ A
j

 

 

' V

96.0 92.0 88.0 84.0 80.0

Binding Energy , eV
Figure 4-1. XPS spectrum for the 41‘5/2 and 4f7/2 levels of Au:(A)

100.0

Au particles from AuCl3 reduction: (B)AuZn particles
formed from reduction of a mixture of AuCl3 and
excess anzz (C) AuZn + Au particles formed from

reduction ofa mixture of AuCl3 and anz with the former

in excess.

67

Table 4-1. Electron diffraction of AuZn from co-reduction of

 

 

 

 

 

 

 

11001341112,

D(cm)a d(A)b d'tAic AuZn(A)d hkle
1 3.31 3.13 3.19 3.1960 100
2 4.63 2.24 2.28 2.2599 110
3 5.73 1.81 1.85 1.8452 111
4 6.58 1.57 1.60 1.5980 200
5 7.40 1.40 1.43 1.4293 210
6 8.14 1.27 1.30 1.3048 211

 

 

 

 

 

 

 

 

a. Ring diameter of electron diffraction.(camera length - 140cm)
b. d-spacing converted from ring diameter.

c. d-spacing corrected by camera constant.

(1. d-Spaclng from powder X-ray diffraction files.(30-608)

e. Miller index.

68
(ll) AU-Cu

The Au-Cu alloy is the most studied of all alloys except Fe-C. It
has been found that the structure at its ideal composition. AuCu.
consists of alternate layers of Au and Cu atoms parallel to a cube
face. The symmetry is slightly distorted to tetragonal with c/a-0.93.
The structure transforms to orthorhombic with b/a-10.03 at about
380 “062) We prepared a new phase of this compound with a cubic
structure as identified by the SAD pattern by reduction of
stoichiometric mixtures of AuCl3 and CuClz at -50 'C.

Formation of a compound in the Au-Cu system was suggested
but not proven by the XPS data. The Au 4f7/2 and 4f5/2 XPS lines
are near those of pure Au. and the Cu 2p3/2 XPS line at 932.3 eV is

close to that of Cu (932.4 eV) and CuzO (932.2 eV). However. the
modified Auger parameter. 01'. which is free of charging errors. is
1849.9 eV. about 1.6 eV below those of copper metal and CuO but
close to the value for CuzO (1849.5 eV). Thus. the photoelectron and
Auger data are not definitive in this case. The SAD results are.
however. clear and unambiguous. The diffraction pattern is of the
same type as that of AuZn and corresponds to a simple cubic
structure with each Au atom at the center of a cube of Cu atoms and
vice versa. The patterns of the two known higher temperature
phases of AuCu (tetragonal and orthorhombic) would be very
different from that observed. The cubic unit cell has a-2.95 A
compared with 3.196 A for AuZn. The d-spacings are listed in Table
4-2.

 

Table 4-2. Electron diffraction of AuCu from co-reduction of

69

 

 

 

 

 

 

 

 

 

 

 

 

AuCl3+CuClz,

Dtcmia 01A)b 0'18)C mad
1 3.60 2.88 2.94 100
2 5.12 2.02 2.06 110
3 6.26 1.66 1.69 111
4 7.27 1.43 1.46 200
5 8.12 1.28 1.31 210
6 8.95 1.16 1.18 211

 

 

a. Ring diameter of electron diffraction.(camera length - 140cm)
b. d-spacing converted from ring diameter.

c. d-spacing corrected by camera constant.

d. Miller index.

70
(111) Tellurides

When particles of Au. Cu or Te were separately prepared by
reduction of AuCl3. CuClz and TeBr4 respectively. the XRD pattern
always showed peaks of the metal. Reduction of a mixture. however.
gave X-ray amorphous products except when an excess of one
component was used. This suggests the formation of alloys or
compounds with either random occupancies or crystallite sizes
smaller than 3 nm.

The reduction of a mixture of CuClz and TeBr4 in Me20 yielded
a powder that was washed with air-free methanol without significant
oxidation. The Cu 2p3/2 and 291/2 binding energies from the XPS of
the product. 932.5 and 952.3 eV respectively. are indistinguishable
from those of metallic copper. and the L3 VV Auger peak is at nearly
the same energy as for pure Cu. The modified Auger parameter. at“.
1850.5 eV. is slightly lower than the average literature value of
1851.5 eV. Thus. the XPS data for the binary system Cu—Te do not
distinguish between compound formation and mixture of the pure
metals. They do show. however. that the original compounds were
reduced.

The case for compound formation in the Au—Te system is
somewhat stronger. The Au 41'7/2 and 4f5/2 XPS peaks in the
binary system are nearly twice as wide as for pure Au. although the
peaks positions are essentially the same as those of the pure metal.
Also the value of 01' for Te in the mixture (1064.5 eV) is at least 0.4

eV lower than that of elemental tellurlum.

 

-L“ .L I (ruJI-‘ILIF-I
l

71

On the basis of the XRD results and the XPS data. it is concluded
that alloy or compound formation occurs in the Au-Te system and

probably also in the Cu-Te system.

(IV) Ti-Au
The reduction of a mixture of TiC14 and AuCl3 yielded the XPS

results shown in Figure 4—2. The unwashed product (top curve) gave
broadened Au 41‘7/2 and 41‘5/2 XPS peaks at the positions observed
with pure gold. (83.7 and 87.4 eV) and shoulders at 84.8 and 88.4
eV. Upon washing with methanol. the XPS pattern (bottom curve)
reverted to that of pure Au. The possible formation of a Ti-Au
compound is suggested by these results. but further work needs to
be done to verify this conclusion. The results could be significant for
catalysis if it were possible to produce small noble metal particles on

a finely divided T102 or Ti(OR)2 support.

72

 

 

96.0 92.0 88.0 34,0

Binding energy, eV

Figure 4-2. XPS of the reduction product of a mixture of TiCl4 and
AuCl3 in MezO (top) and after washing with air-free

methanol (bottom).

 

73
(V) Ta-Fe

Refractory and oxophilic metals. such as Ta. W and Mo present
severe challenges to preparation by this method. They require
strong reductants. have multiple oxidation states and have such a
high affinity for oxygen that small particles are expected to be very
reactive. The “intertness” of the bulk metals is due to a compact
oxide layer rather than to an inherent lack of reactivity. In spite of
these difficulties. our initial results showed that Mo and W metals
could be produced by reduction of MoCl5 and WC16 respectively. but.
as expected. washing with methanol resulted in complete oxidation.

Considerable effort has been expended in an attempt to
prepare millimole quantities of small tantalum particles for use as
precursors in ternary nitride synthesis. Purified TaC15 is very
soluble in Mezo and reacts immediately in an H-cell with K*(15C5)2K'
prepare in situ. Since five moles each of K and 15C5 are required per
mole of TaC15. copious amounts of KCl and K*(15C5)2Cl‘ are produced
and are only slightly soluble in MezO. Most. but not all. of the by-
products can be removed by washing with anhydrous liquid
ammonia. The product is Ta metal together with some unidentified
organic material. The fine black powder is ‘pyrophoric and reacts
immediately with water and methanol. It can. however. be handled
and stored in an inert atmosphere box without apparent change. The
XRD pattern shows the product to be amorphous. but the XPS peaks
(Figure 4-3-D) provide identification as Ta metal.

Pyrophoric iron can be readily prepared by reduction of FeCl3
in Mezo. When solid FeCl3 was present in contact with the saturated

solution during reduction with an excess of K*(15C5)2K'. crystalline

 

 

74

Fe was produced as indicated by the XRD pattern. The by-products
could be removed by washing with liquid ammonia. More dilute
solutions yielded X-ray amorphous iron. The XPS peaks (Figure 4-3-
B) show the presence of both metallic iron and oxidized iron on the
surface of the particles.

In an attempt to prepare the known compound FezTa. mixtures
of FeCl3 and TaC15 in a 2:1 mole ratio were reduced with K*(15C5)2K‘
in MezO and the product was washed with NH3. The product has not
yet been identified. but the XPS data show the presence of Ta. Fe and
K in approximately a 1:2:3 mole ratio. The Fe and Ta XPS patterns
are shown in Figures 4-3-A and 4-3—C respectively. The latter shows
a shift of the Ta 41’7/2 peak from 21.6 eV in Ta metal to 23.8 eV.
Oxidation of the mixture in air shifts the peak to 24.8 eV. The Fe
2p3/2 peak at 706.3 eV (Fe metal - 706.8 eV) shifts to 709.7 eV upon
oxidation (Fe203 - 711 eV). The K 2p3/2 peak at 292.6 eV does not
shift upon oxidation. A substantial carbon is peak at 285.0 eV
decreases dramatically upon argon ion sputtering. The results show
that some type of complex alloy or compound formation occurs in the
K-Ta-Fe system. but the product or products have not been
identified.

75

.25: mag :58 2x 69 R: ”5.3.
2 scan 3 EB. Ex .6... 0V 8:26 906m .8: max on EV ”20$ 2 56". 3 53
max on A5 ”mmz 256.25 53, 62.33 .2605 ”839?. 9%.? 65 Ba 3:52 max mtv 2:3"—

.>o 33.35 25...:—

 

 

fiu 38 6.3 33 32. 3:. 9g.
/.. . //1) -1 1.1.1111
./ \ :
,./ \x .\ u “u “z
/,

76

C. Summary

In spite of the preliminary nature of these experiments. it is
clear that reduction of pure compounds in MeZO or THF solutions
with dissolved alkalides and electrides leads to nanoscale metal
particles. The strong reducing ability of alkalides and electrides
makes the method very general and permits metal production across
the transition metal and post-transition metal series. from Ta to Te.
It is difficult to positively identify alloys and compounds with
particle sizes too small to yield an X-ray diffraction pattern and with
reactivities too high to load into available TEM without oxidation.
While the less reactive compounds AuCu and AuZn could be
characterized. a number of other binary systems yielded powdered
samples whose XPS pattern suggested alloy or compound formation.
but did not provide identification. This method is also applicable to
the formation of finely divided metals on oxide supports and to the

synthesis of organometallic compounds.

 

Chapter 5 REDUCTION in ZEOLITE PORES

A. Introduction

In recent years there have been considerable academic and
industrial research efforts carried out in the field of zeolite catalysis.
There are 34 known natural zeolites and about 100 zeolites which do
not have natural counterparts have been synthesized. Of this large
number of zeolites. only a few have found commercial application:
they are mostly synthetic zeolites and synthetic-analog natural
zeolites. A zeolite has been defined as a “crystalline aluminosilicate
with a tetrahedral framework structure enclosing cavities occupied
by cations and water molecules. both of which have enough freedom
of movement to permit cation exchange and reversible dehydration”.

The major part of zeolite catalysis work has been related to
reactions where the zeolite is used as solid acid. e.g.. isomerization.
cracking. hydrocracking etc. The catalyst of choice for catalytic
cracking. the heaviest use of zeolites. is usually a rare earth.
magnesium. hydrogen. or ultrastable form of zeolite X and Y. or a
combination of these.

In general. zeolites have four properties that make them
especially interesting for heterogeneous catalysis:(53)

1. They have exchangeable cations allowing the introduction of
cations with various catalytic properties.
2. If these cationic sites are exchanged with 11*. they can have a

very high number of very strong acid sites.

77

 

78

3. Their pore diameters are less than 10 A.
4. They have pores with one or more discrete sizes.

The properties of a zeolite are dependent on the topology of its
framework. the size. shape. and accessibility of its free channels. the
location. charge and size of the cations within the framework. and the
presenCe of faults and/or occluded material. Therefore. structural
information is extremely important in understanding the adsorptive
and catalytic properties of zeolite catalysts.

The fundamental building block of all zeolites is a tetrahedron
of four oxygen anions surrounding a small silicon or aluminum ion.
These tetrahedra are arranged so that each of the four oxygen anions
is shared in turn with another silica or alumina tetrahedron. The
crystal lattice extends in three-dimension and the -2 oxidation state
of each oxygen is accounted for. Each silicon ion has its +4 charge
balanced by the four tetrahedral oxygens and the silica tetrahedra
are electrically neutral. Each alumina tetrahedron has a residual
charge of -1 since the trivalent aluminum is bonded to four oxygen
anions. Thus each alumina tetrahedron requires a +1 charge from a
cation in the structure to maintain electrical neutrality.

The silica and alumina tetrahedra are combined into more
complicated “secondary units. which form the building blocks of the
framework zeolite crystal structures. The silica and alumina
tetrahedra are geometrically arranged. with Al-O-Al bonds excluded.

The unit cell formula is usually written as

M“*x/nl(A102‘)x(Si02)yI szO

 

79

where M“* is the cation which balances the negative charge
associated with the framework aluminum ions. These metal cations.
which neutralize the excess anionic charge on the aluminosilicate
framework. are usually alkali metal and alkaline earth metal cations
and at least some of them must be able to undergo reversible ion
exchange. Water molecules fill the remaining volume in the
interstices of the zeolite.

The tetrahedra are arranged so that the zeolites have an open
framework structure. which defines a pore structure with a high
surface area. The three-dimensional framework consists of channels
and interconnected voids or cages. The cations and water molecules
occupy the void spaces in the structure. The intracrystalline zeolite
water can be removed by thermal treatment. usually reversibly.

Zeolite Y consists of tetrahedra linked to form so-called sodalite

cage units as shown in Figure 5-1.

A7

\ .
___d

 

 

Figure 5-1 Structure of sodalite cage
A sodalite unit is the secondary building block of zeolite Y. Molecules
can penetrate into this unit through the six-membered oxygen rings.
which have a free diameter of 2.6 A: the unit contains a spherical
void volume with a 6.6 A free diameter. Since the pore diameter is
so small. only very small molecules. e.g. water. helium. or hydrogen

cations can enter the sodalite cage.

 

80

When the truncated octahedra are connected by bridge oxygen
atoms between the six-membered rings. zeolite Y is formed. Figure
5-2 shows the three-dimensional structure of zeolite Y. The unit cell
of zeolite Y is cubic with a unit cell dimension of 25 A. and it contains
192 silica and alumina tetrahedra. The unit cell dimension varies
with Si/Al ratio. Each sodalite unit in the structure is connected to
four other sodalite units by six bridge oxygen ions connecting the
hexagonal faces of two units. The truncated octahedra are stacked
like carbon atoms in diamond. This structure results in a supercage
surrounded by ten sodalite units which is sufficiently large for an
inscribed sphere with a diameter of 12 A. The opening into this
large cavity is bounded by sodalite units. resulting in a 12-
membered oxygen ring with a 7.4 A free diameter. Each cavity is
connected to four other cavities. which in turn are themselves
connected to three-dimensional cavities to form a highly porous

framework structure.

 

 

 

 

 

 

Figure 5-2 Structure of zeolite Y

 

81

Molecular sieve zeolites have played an important role in
catalysis since the 1960's because of their selective adsorptive
properties.(54) Their properties have been reviewed recently.( 55)
In order to locate and disperse metal particles in zeolite. researchers
have studied Zeolites intensively by X-ray diffraction. small angle X-
ray scattering. chemisorption. electron microscopy. IR spectroscopy.
X-ray photoelectron spectroscopy and Auger electron spectroscopy.
The zeolite Y occurs naturally as the mineral faujasite and consists of
a porous network of aluminate and silicate units tetrahedrally linked
through bridging oxygen atoms. The structure consists of sodalite
units arranged in a diamond net and linked through double six-
rings.(56) These cavities provide an ideal environment for
synthesizing single-size clusters.

The traditional way to synthesize supported metal zeolites is
by impregnation or ion exchange followed by hydrogen reduction to
give zero valent metal. In most cases. the final dispersion and
location of the metal depend upon the pre-treatment under oxygen.
which governs the cation positions. and the gaseous environment in
which the sample is thermally treated before reduction.
Homogeneous reduction of various metal salts with alkalides or
- electrides in dimethyl ether or tetrahydrofuran at - 50 'C or below
has produced 3- 15 nm diameter particles of metals. alloys or
oxidized products as described in Chapters 3 and 4. Here. the new
reduction 'method that uses alkalides and electrides in zeolite pores

will be discussed.

 

82

B. Experimental

One gram of NaY zeolite. Na56 (A102)56(8102)136XH20. was ion
exchanged by 100 m1. of 0.1 N AgNO3. or NadeCl4. or CuClz in
aqueous solutions for 3 days. and was washed by H20. methanol. and
dry ether several times to remove excess ions. followed by vacuum
drying for 3 days at 120'C. AgY. PdNaY and CuNaY were produced
respectively. About 200 milligrams of ion-exchanged zeolite and an
excess of electride. K*(15C5)2e‘. were added in a nitrogen-filled
glove bag to different sides of an evacuated H-cell (Figure 2-1) with
a medium frit. A liquid nitrogen bath was used to cool the cell to
prevent decomposition of the electride. About 20 mL of pre-purified
MezO was distilled into both sides after evacuation to about 10'5
torr while the cell was kept in a -50 'C isopropanol bath.

The blue-black solution of electride was poured through the
frit to react with zeolite. Reaction occurred immediately upon
addition of the electride as indicated by the fading of the blue color.
A slight excess of electride was added until the blue color no longer
disappeared to make sure that the reaction was complete. Then the
solvent was distilled out under vacuum. Pre-purified water or
methanol was used to wash away the by-products. The zeolite was
removed from the H-cell and mounted on either indium foil or 3M
Scotch double-coated tape in the He-filled dry box. A vacuum
transfer vessel was used to carry the sample from the dry box to the
XPS transfer chamber without exposing the sample to air.

The X-ray photoelectron spectra were obtained by using

monochromatic Al K01 X-rays. The Si 2p line at 102.8 eV was used as

 

83

an internal standard. The charging effect was reduced by flooding
the sample with zero kinetic energy electrons of a neutralizer. X-ray
diffraction patterns were recorded by packing the zeolite on a glass
slide. A drop of zeolite suspension was put on the transmission

electron microscope carbon coated grid and allowed to dry in the air.

C. Results

The XPS spectra of palladium in zeolites are shown in Figure 5-
3. The reduction of palladium containing zeolite in dimethyl ether by
electride at -50 '0 leads to a shift of the 3d5/2 line position from
337.5 eV to 335.2 eV: this is the evidence for the reduction of
palladium cations. This binding energy of palladium in zeolite is the
same as that of a metallic palladium film.(57) However the value of
the binding energy of the Pd 3d5/2 level in zeolite Y after hydrogen
reduction is 336.1 eV as determined by Romannikov and co-
workers.(58) Previous work had shown that atomically dispersed
Pd' in zeolite Y is at 336.4 eV while 20A crystalline Pd' is at 335.2
eV.(59)

 

84

N(E)

 

e f j I

i

if

xv] I

. \-/~\/‘ ..
E
i

/ 1‘. l" i
/ ‘1
i

350 346 342 338 334

Binding Energy, ev

<>(
4 .—

4 A
f V

Figure 5-3. XPS spectra of Pd in zeolite Y before reduction
3d5/2-337.5 eV. 3d3/2-343.0 eV (top) and after
reduction 3d5/2-335.2 eV. 303/2-3404 eV (bottom).

 

85

We have calculated atomic ratios from the peak areas and the
cross sections reported by Scofield.(60) These values were compared
with the theoretical ratios obtained by assuming no surface
enrichment. Surface enrichment of _a particular element would
normally appear as a large relative increase in the peak for that
element since XPS only detects the top few layers of the zeolite
particles. It is clear that the Pd 4f/Si 2p ratios did not change very

much during reduction. Figure 5-4 shows the survey scans of (A)

plain zeolite Y (B) after NadeCl4 ion exchange ( new peaks of Pd
3d5/2 at 337 eV and 3113/2 at 343 eV appear and the Na ls peak at
1071 eV decreases) (C) after reduction by. K+(15C5)2e'.(new peaks
of K 2p appear showing potassium cations balance the counter
charge)

Powder X-ray diffraction patterns were recorded to make sure
that there was no structural change during reduction.(6l) The peak
corresponding to the (111) d-spacing of palladium was not observed
for the sample. This indicates that the palladium particles are too
small (< 30A) to be detected by powder X—ray diffraction.
Transmission electron microscopic investigation also showed the
absence of metallic particles on the external surface of the zeolite

crystals.

 

86

 

A L
v f

-+—+-+— + ++—<L

“UR
.— 1» u .A a 0- V o o ‘

 

 

I

AAA A A A A A A A A A A A
v V V V T v V T v Y

 

5'.
4
4

 

 

 

KIM
h. N w A u '5 V O O
- - +——+——+—+—+—o—+——o o—+ 4 %
4—-6——6-—¢-—6—0-6—+—6—+++

4.+ 4W1. + +—+—
y
I
J»

 

 

a.
-

E 4
4

i

1‘ .

4»

it

t

J»

‘1

11
- 4 o 4.4—

A AA AA; A A A

 

 

l“ )5
o - N u .A u a V a u

1)
3
.:

 

Figure 5-4. XPS survey scans of (A) plain zeolite Y (B) after Na2PdCl4
ion exchange ( new peaks of Pd 3d5/2 at 337 eV and
3113/2 at 343 eV appear and Na ls peak at 1071 eV
decreases) (C) after reduction by. K+(15C5)2e’.(new
peaks of K 2p appear showing potassium cations balance

the counter charge)

 

87

Copper cations in Zeolite Y were reduced by electrides as shown
in Figure 5-5. The binding energy of the copper 2p3/2 level shifts
from 934.35 to 933.70 eV after reduction. The peak was shifted 2.1
eV to higher binding energy after eXposure to air for 24 hours. We
assume that this is the oxidation of copper to cupric oxide in zeolite
because the peak maximum is rather poorly defined with respect to
the Cu' peak due to significant broadening. probably associated with
multiple splitting. Moreover. the strong shakeup peak located 10 eV
above the principle copper 2p3/2 line is a good indication of the
presence of CuO.(62) The copper L3M45M45 Auger spectra of Cuzfi
Cu0 and CuO in zeolite are all shifted to different kinetic energies.
However. these peaks are broad and somewhat more difficult to
characterize because of the large number of. discrete Auger lines
under this envelope. There is no surface enrichment of copper either
after reduction or re-oxidation. Transmission electron microscopy
and powder X-ray diffraction were also obtained. Clean surfaces and

no structural change of the zeolite were observed.

88

 

 

+~++++ -+—+——+—+—+——+-+—+-—+~+—+—¢—t——+——+—+—r—o—+—¢—+—+

I
1

A A J—AA A A A A .A A P A #9 A
V7 Y 1 vi

970: 960 960' ' 940 930
Binding Energy. ev

Figure 5-5. XPS spectra of Cu in zeolite Y before reduction (2p3/2-
934.35 eV. 2p1/2-954. 15 eV top). after reduction
(2p3/2-933.7 eV. 2p1/2-953.8 eV middle). and oxidized
in the air (2p3/2=935.8 eV, 2p1/2=955.9 eV bottom).

89

Silver cations were reduced by electrides in Me20. The silver
3115/2 peak shifted from 369.5 to 368.75 eV after reduction with
great surface enrichment. Figure 5-6 shows XPS spectra of Ag in
zeolite Y before reduction 3d5/2-369.5 eV. 3d3/2-375.6 eV (top)
and after reduction 3d5/2-368.7 eV. 3d3/2-374.7 eV (bottom).
Figure 5-7 shows the survey scans of (A) plain zeolite Y (B) after
AgCl ion exchange (new peaks of Ag 3115/2 at 369 eV and 3113/2 at
375 eV appear and Na ls peak at 1071 eV disappear) (C) after
reduction by . K*(15C5)2e'.(new peaks of K 2p appear showing
potassium cations balance the counter charge). The average particle
size estimated by Scherrer's equation from the powder X-ray peak
broadening is about 40 A. A TEM micrograph of reduced silver
zeolite also confirmed that silverparticles were on the external
surface of the zeolite crystals. The reduced silver clusters may
migrate out of zeolite cavities after reduction and aggregate to larger
and stable particle size. Alternatively. the silver cations may
exchange with potassium cations first and then react with electride

on the surface of the zeolite.

 

90

393/2 3d5/2 '1‘“

AAAAAJAAA
I roV I it V

 

 

N(E) \_MAW

\
\ 1

.. ‘\

. *w”

 

1r—+—+ f ‘+-+
J
A

J A 4‘
vi Y

7382‘ I 378' ' 8374* ' 370' ' 366'
Binding Energy. eV

Figure 5-6. XPS spectra of Ag in zeolite Y before reduction
3d5/2-369.5 eV. 3d3/2-375.6 eV (top) and after
reduction 3d5/2-368.7 eV. 3d3/2-374.7 ev (bottom).

91

 

 

I“ )1!

 

 

 

 

 

I“ )R

 

 

 

I“ )ll

1
‘b I
o y WW+»+~4€4+—+—++o o o _+,.,__,_+_+_+

 

 

- H N H ‘ h” a
A A A A A A A A A
V V

 

 

' Figure 5-7 XPSsurvey scans of (A) plain zeolite Y (B) after AgNO3 ion
exchange (new peaks of Ag 3d5/2 at 369 eV and 3113/2
at 375 eV appear and Na ls peak at 1071 eV disappear)
(C) after reduction by . K*(15C5)2e‘.(the peak intensities
of Ag show surface enrichment and new peaks of K 2p

appear showing potassium cations balance the counter

charge)

92

D. Summary

Electrides. the most powerful reducing agents. can reduce Cuzfi
Pd2*. and Ag+ originally present in Y zeolite pores. XPS data show
that these cations were completely reduced to the zero oxidation
state by K+(15C5)2e‘ at -50 'C. XPS is clearly a valuable tool for
studying small metal particles or catalysts. In addition to its surface
sensitivity. XPS is uniquely capable of providing chemical
information such as oxidation state and chemical bonding. as well as
elemental compositions.

XPS shows that the electride is powerful enough to reduce Pd .
Cu and Ag cations to the metallic state. XRD data indicates that there
is no structural change of the zeolite during the ion-exchange and
reduction precesses. A TEM investigation showed no particle
aggregation on the zeolite surface in the cases of Pd 81 Cu. for which it
is believed that Pd and Cu were reduced to the metallic state and
remained trapped inside the zeolite pores. while K+ balances the
counter charge upon reduction. However. Ag formed particles that

were 40 A in diameter. uniformly dispersed on the external surface.

Chapter 6 CONCLUSIONS and SUGGESTIONS for FUTURE WORK

The science and technology of nanometer particles have been
greatly advanced to meet the urgent demands of modern industries
including catalysts. sensors. electromagnetic devices. and
photosensitive materials. Monodispersed colloidal systems are
invaluable for this purpose because the entire system‘s uniform
physical/chemical pr0perties directly reflect the properties of each
individual particle. The highly dispersed solid state is
thermodynamically unstable. thus highly dispersed metals are not
found in nature. However. metals with high melting temperature can
be obtained and maintained in this state over long periods of time.
since aggregation involves a high activation energy.

The obiective of this study was to test the generality of
homogeneous reductions of soluble metal salts by solvated electrons
and/or alkali-metal anions in aprotic solvents. Although many of the
metals studied can be prepared as small particles by other methods.
this method has the advantage of being rapid and quantitative and of
requiring only low temperatures for all steps. The initial reduction
products must consist of very small. reactive particles as binary alloy
formation is observed rather than mixtures of metals when two
metal salts are simultaneously reduced. This new method for the

preparation of metal/oxidized metal or alloy particles not only

93

 

94

provides the nanoscale particles of metals and alloys but also
produces relatively uniform particle size.

For a single element reduction. soluble compounds of transition
metals and post-transition metals in dimethyl ether or
tetrahydrofuran are rapidly reduced at -30 'C by dissolved alkalides
or electrides to produce metal particles with crystallite sizes from < 3
to 15 nm. The products were characterized by X-ray photoelectron
spectroscopy. powder X-ray diffraction and selected area electron
diffraction The average particle size was estimated from the line
broadening of powder X-ray diffraction. Particle size distributions
were determined by counting the particles on electron micrographs
obtained by transmission electron microscopy. Salts of Au. Pt. Cu. Te.
Fe. and Ta formed metallic particles with little or no oxidation even
when washed with methanol or liquid ammonia. Reduction of salts
of Ni. Fe. Zn. Ga. Si. Mo. W. In. Sn. and Sb yields surface oxidation over
a metallic core. This method is also applicable to the formation of
finely divided metals on oxide supports.

X-ray amorphous alloys or intermetallic compounds formed
when two different metal salts were used. as indicated by X-ray
photoelectron spectroscopy. Selected area electron diffraction
revealed the structure of air-stable samples. Reduction of salts of
Au-Cu. Au-Zn. Cu—Te. Zn-Te. Au-Ti. Fe-Ta formed particles that
contained intermetallic compounds.

The reduction of Cuz”. sz“. Ag+ in zeolite-Y was made. X-ray
photoelectron spectroscopy data show that these cations were
completely reduced to zero oxidation state at -30 'C. Cu and Pd

remain trapped in the interior and an alkali metal cation balances

 

95

the counter charge upon reduction. In contrast. Ag formed 40 A
diameter particles uniformly on the external surface.

The nanoscale metal particles are extremely air-sensitive: they
will be oxidized in a few seconds without any protection in the air.
X-ray photoelectron spectroscopy has proven to be a useful tool in
the study of nanoscale particles. In addition to its surface sensitivity.
XPS is capable of providing chemical information such as the
oxidation state and the nature of the chemical bonding. as well as
elemental composition.

Powder X-ray diffraction and selected area electron diffraction
provide additional characterization of those air-stable particles.
Average particle size was estimated from the line broadening of
powder X-ray diffraction. Particle size distributions and morphology
was made from the electron micrographs of transmission electron
microscopy.

The small angle X-ray scattering has undergone considerable
progress in recent years. Experiments on the characterization of
supported metallic catalysts have demonstrated that even for low
metal concentration(0.67. Pt on alumina). distribution curves of the
particle diameters can be obtained which are comparable to electron
microscopy results.(63) Moreover. one of the main advantages of
this technique is the possibility to obtain directly a parameter
proportional to the metallic surface area.

This technique is especially favorable for the preparation of
metal particles in zeolites. As the pores are regularly arranged in the
zeolite. the scattering produced by the support is very weak.

Therefore about 95 74 of the small angle X—ray diffraction intensity is

96

related to metal. Clearly this technique can give one a true particle
size distribution over the pure metallic particles or metal particles on
support. I think the small angle X-ray scattering should be the next
step to use. ‘

Extended X-ray absorption fine structure (EXAFS) is another
method that can give insight into the structure and dimensions of the
small metallic particles. This method is based on absorption of X-
rays resulting from the excitation of an electron from core states and
the subsequent scattering of the outgoing photoelectron waves by
neighbors of the excited atom. Experimentally. the method requires
intense X-rays sources and thus became practicable only recently
due to the availability of intense synchrotron X-ray sources.

As a structural technique. the EXAFS method has many
advantages. Since the X—ray absorption is specific for a given
element. it is possible to analyze separately the environment of each
kind of atom in a multi-component system. Due to its sensitivity to
the immediate neighborhood of the X-ray absorbing atom. a single
crystal is not required and the method can be successfully used both
with small size crystallites and even amorphous solids. 1n the case of
extremely small microcrystals and clusters. the mean coordination
number around a given atomic species is smaller than that observed
in massive solids.

Due to these advantages. 3 significant amount of work has
already been performed on highly dispersed metals with
investigations of the local chemical composition. structure and
determinations of the size distributions of the smallest clusters. In

particular. supported platinum and copper catalysts have been

 

97
investigated with this technique.(64. 65) EXAFS should be another

technique for the characterization of small particles and supported
particles.

Colloidal particles are normally formed through a sequential
process of nucleation and growth of the nuclei. To obtain a
monodispersed system. the two stage must be strictly separated and
nucleation avoided during the growth period. Since the steady
concentration of monomers in the growth stage is determined by the
balance between the rates of generation of monomers and their
consumption by particle growth. the generation or introduction rate
of monomers must be controlled so that it is low enough to keep the
balanced monomer concentration below critical supersaturation after
the nucleation period.(66)

Typically. the initial concentration of metal salts.
alkalide/electride. and temperature must be adjusted to meet this
requirement for the formation of monodispersed particles. 1 have
demonstrated that the particle sizes can be decreased by lowing the
initial concentration of the metal salts. In order to produce
monodispersed small metallic or alloy particles. I think much more
work can be done by controlling the temperature and the
concentration of alkalide/electride.

The reduction with alkalides/electrides is so powerful that
alloys can be formed instead of separated metals when mixed metal
salts are reduced. A lot of alloys which are very useful in industrial
applications might be made by this method including:

semiconductors (GaAs). shape memory alloys (NiAl). magnetic

particles (Fe203. NiFe203. and CoFe204). It is hope that this

 

98

introduction of a new low temperature method for the preparation of
nanoscale metal and alloy particles will lead to further developments

by others.

 

APPENDIX

 

compound

trials

comments

 

AUCI3

9

The solubility of AuCl3 is high both in
Me20 and THF. Colloidal Au solution
was formed first. The Au particles
start to aggregate a few seconds later
in Me20. The colloidal Au is stable in
THF at least 3 hours. The black Au
particles can be washed by MeOH or
H20 in the air easily.

 

CuClz

15

The red colloidal Cu is stable in THF for
at least 48 hours in room temperature.
The black Cu products can be washed
by MeOH under protection. The
unwashed products turned green
several days later in the air.

 

TeBr4

The solubility of TeBr4 is high both in
MeZO and THF. The black metallic Te
can be collected by washing the
reaction products with MeOH under
protection.

 

anz

24

The Zn product will be oxidized easily
without careful protection. Only
broadened ZnO XRD lines were
observed from MeOH washed products.

 

NIClz

Triethyl phosphine was added to
increase the solubility of NiClz in Me20
and THF. The Ni products can be
oxidized easily without good
protection. Ni(Et3P)zC12 can be reduced
directly by alkalide or electride in
Mego or THF.

 

 

 

FeCI3

 

 

A deep green colloid was formed when
washed with MeOH. The clean metallic
Fe can be collected by washing with
NH3 in the H-cell.

 

99

 

100

 

63C] 3

Both metallic Ga and oxide peaks were
observed for unwashed and MeOH
washed products from XPS.

 

SIC14

Both elemental Si and oxide peaks
were observed for unwashed and
MeOH washed products from XPS.
Only oxide peaks were observed if the
products were not carefully protected.

 

GeCI4

Both metallic Ge and oxide peaks were
observed for unwashed and MeOH
washed products from XPS.

 

10

SbCl5

Both metallic Sb and oxide peaks were
observed for unwashed and MeOH
washed products from XPS.

 

11

5110/,

Both metallic Sn and oxide peaks were
observed for unwashed and MeOH
washed products from XPS.

 

12

InCI3

Both metallic In and oxide peaks were
observed for unwashed and MeOH
washedproducts from XPS.

 

13

MOC15

Both metallic Mo and oxide peaks were
observed for MeOH washed products
from XPS. Only Mo oxide was
observed upon washing by MeOH.

 

I4

WC16

Both metallic W and oxide peaks were
observed for MeOH washed products
from XPS. Only W oxide was observed
upon washing by MeOH.

 

l5

H2PtC16

Black metallic Pt can be washedby
MeOH or H20. The particle size is
about 3-4 nm from TEM microgragh.

 

l6

TaCI 5

Ta metallic particles can be washed
cleanly by liquid NH3. Only Ta oxide
was observed upon washing by MeOH

 

l7

MnI2

Products were unable to be identified
at this time.

 

 

18

 

AICI3

 

 

The unwashed reduction products
were not identified by XPS because the
peak separation between metallic Al
and oxide is too narrow. Gas evolution
was observed upon washing with
MeOH.

 

 

101

 

 

 

 

 

 

 

 

 

 

 

19 VC13 4 VC13 was reduced by alkalide or
electride. The products have not yet
been characterized yet.

20 Cd]; 2 Metallic Cd was observed in the
unwashed reaction products. The
oxidized product was identified as
Cd(OH)2 by XPS upon washing with
MeOH.

21 YCl3 2 Only Yittrium oxide was observed in
the unwashed products.

22 TIC14 11 Only Titanium oxide was observed in
the unwashed products and MeOH
washed products.

23 AuCl3 5 The products were identified as AuZn

Zn]; (CsCl type structure) by XPS and SAD.

24 AuCl3 4 The products were identified as Cu Au

CuClL (CsCl type structure) by SAD.
25 AuCl3 3 Two new peaks of Au 415/2 and 4f7/2
T114 from XPS shows that there is a new
compound formed. The compound has
not yet been identified.

26 AuCl3 3 Reduction of this mixture gave X—ray

TeBr4 amorphous products except when an
excess of one component was used.
XPS data suggested that a new
compound was formed.

27 CuC12 3 Reduction of this mixture gave X-ray

TeBr4 amorphous products except when an
excess of one component was used.
XPS data suggested that a new
compound was formed.

28 CuClz 2 Reduction of this mixture gave X-ray

anz amorphous products except when an

 

 

 

excess of one component was used.
XPS data suggested that a new,
compound was formed.

 

 

 

102

 

 

 

 

 

 

 

 

 

 

 

 

29 TaC15 The product has not yet been
FeCl3 identified. but the XPS data show the
presence of Ta. Fe and K in
approximately a 1:2:3 mole ratio when
a 1:2 ratio of TaC15 and FeCl3 was
reduced by excess K*(15C5)2K‘ in
M820.
30 SbC15 Unable to identify the products at
GaC13 present time.
31 AlCl3 Unable to identify the products at
TeBr4 present time.
32 Cd]; Unable to identify the products at
TeBr4 present time.
33 CuClz Unable to identify the products at
MnClzr present time.
34 InCl3 Unable to identify the products at
TeBr4 present time.
35 AuCl3 Unable to identify the products at
GaC13 present time.
36 Cu2+ Cu2+ loading percentage can be
on Na-Y increased by multiple step ion-
exchange. Metallic Cu peaks were
observed. These Cu particles may stay
inside the sodalite cage or the super
cage of the zeolite-Y. The metallic Cu
clusters were oxidized easily upon
exposure to air.
37 Pd2+ PdZ+ loading percentage can be
on Na-Y increased by multiple step ion-
exchange. Metallic Pd peaks were
observed. These Pd particles may stay
lnside the sodalite cage or the super
cag of the zeolite-Y.
38 Ag* Almost 100% Na+ was replaced by Ag+
on Na-Y after ion exchange. Metallic Ag
particles were observed on the surface
of zeolite-Y.
39 AuCl3 Au3+ did not ion-exchange with Na+ in
on Na-Y zeolite—Y. Au particles on the surface

 

 

 

of zeolite-Y were observed after
reduction with alkalide or electride.

 

 

 

 

103

 

 

40 AuCl3 Au particles with 4 nm diameter were
on A1203 observed on the surface of A1203 from
TEM micrgragh.
41 HthC16 No individual Pt particles were
on A1203 observed on the surface of A1203 from

 

 

 

TEM microgragh. but Pt signal was
observed by EDS and a metallic Pt
pattern was observed by SAD.

 

 

 

(1)

(2)

(3)

(4)

(5)

(6)

(7)
(8)

(9)

(10)

(ll)

(12)
(13)

References

C. W. Weyl. Annalen 11er Physik and Chemie. 197. 601 (1864).
C. J. Pedersen J. Am. Chem. Soc.. 89 . 7017 (1967).

B. Dietrich. J. M. Lehn and J. P. Sauvage. Tet. Letters. 34. 2855
(1969).

B. Dietrich. J. M. Lehn and J. P. Sauvage. Tet. Letters. 34. 2889
(1969).

J. L. Dye. J. M. Ceraso. M. T. Lok. B. L. Barnett. and F. J. Tehan. J.
Am. Chem. Soc.. 96. 608 (1974).

A. S. Ellaboudy. P. B. Smith and J. L. Dye. J. Am. Chem. Soc.. 105.
6490 (1983).

J. L. Dye. Angew. Chem. Int. E11. Eng. 18. 587 (1979).

J. L. Dye. C. W. Andrews and J. M. Ceraso. J. Phys. Chem. 79.
3076 (1975).

R. S. Bannwart. S. A. Solin. M. G. DeBacker. and J. L. Dye. J. Am.
Chem. Soc.. 111.5552 (1989).

D. lssa. A. E. Ellaboudy. R. Janakiraman. and J. L. Dye. J. Phys.
Chem. 88.3847 (1984).

S. B. Dawes. D. L. Ward. R. H. Huang. and J. L. Dye. J. Am. Chem.
Soc.. 108. 3534 (1986).

S. Doeuff. K. L. Tsai. and J. L. Dye. Inorg. Chem. 30. 849 (1991).
J. S. Landers. A. Stacy. M. J. Sienko. and J. L. Dye. J. Phys. Chem..
85.1096(1981).

104

 

(14)

(15)

(16)

(17)

(18)

(I9)

(20)

(21)

(22)

(23)

(24)

(25)

(26)

(27)

(28)
(29)

(30)

105

J. Papaioannou. S. Jaenicke. and J. L. Dye. J. Solid State Chem.
67. 122 (1987).

K. J. Moeggenborg. J. Papaioannou. and J. L. Dye. Chem. Mater..
3.514(1991). .

J. L. Dye. Scientific American. 257. 66 (September. 1987).

J. L. Dye. J. Phys. Chem. 84. 1084 (1980).

F. Fievet. J. P. Lagier and M. Figlarz. MRS Bull.. December. 29
(1989).

J. Livage. M. Henry. J. P. Jolivet and C. Sanchez. MRS Bull..
January. 18 (1990).

J. J. K. Schloten. J. A. Konvalinka and F. W. Beekman. J. Catal..
28. 209 (1973).

R. M. Wilenzick. D. C. Russel. R. H. Morris and S. W. Marschall. J.
Chem. Phys. 47. 533 (1967).

D. W. MacKee and F. J. Norton. J. Phys. Chem. 68. 481 (1964).
R. D. Rieke. Science. 246. 1260 (1989).

L. Chao and R. D. Rieke. J. Organometal. Chem. 67. C64 (1974).
R. D. Rieke and S. E. Bales. J. Am. Chem. Soc.. 96. 1775 (1974).
R. D. Rieke and P. M. Hudnall. J. Am Chem. Soc.. 94. 7178
(1972).

M. J. McCormick. K. B. Moon. S. R. Jones and T. P. Hanusa. J.
Chem. Soc. Chem. Commun.. 778 (1990).

P. G. Fox. J. Ehretsman and C. E. Brown. J. Catal.. 20. 67 (1971).
W. B. Philips. E. A. Desloge and J. G. Skofronick. Appl. Phys. 39.
3210 (1967).

K. J. Klabunde. H. F. aner. T. O. Murdock and R. Rople. J. Am.
Chem. Soc.. 98. 1021 (1976).

 

(31)
(32)

(33)

(34)

(35)
(36)

(37)

(38)

(39)
(40)

106
J. H. Jean and T. A. Ring. Langmuir. 2. 251 (1986).

H. Bonnemann. W. Brijoux and T. Joussen. Angew. Chem. Int. Ed.
Engl.. 29. 273 (1990).

See. for example. (A) C. D. Wagner. L. H. Gale and R. H.
Raymond. Anal. Chem. 5.466 (1976). (B) D. Briggs and M. P.
Seah. “Practical Surface Analysis by Auger and X-ray
Photoelectron Spectroscopy". John Wiley 81 Sons. Inc.. New
York(1983).

See. for example. (A) A. Schuster and J. W. Nicholson. “An
Introduction to the Theory of Optics”. Edward Arnold and Co..
London. third edition. (1924). pagel 1. (B) H. P. Klug and L. E.
Alexander. “X-ray Diffraction Procedures for Polycrstalline
and Amorphous Materials”. John Wiley 81 Sons. Inc.. New York.
491-583 (1962).

P. P. Edwards. Adv. Inorg. Chem. Radiochem. 25.135. (1982).
J. C. Thompson. Electrons in liquid ammonia.(Oxford University
Press. 1976)

J. J. Lagoski and M. J. Sienko. Eds. “Metal ammonia Solutions”.
(IUPAC. Butterworths. London. 1970).

J. L. Dye. Conference Organizer. Colloque Weyl IV. “Electrons in
Fluids- The Nature of Metal Ammonia Solutions”. J. Phys. Chem.
79. No 26 (1975).

C. A. Kraus and H. F. Kurtz. J. Am Chem. Soc.. 47. 43 (1925).

J. L. Dye. M.G. DeBacker and V.A. Nicely. J. Am Chem. Soc.. 92.
5226 (1970).

 

(41)

(42)

(43)
(44)

(45)

(46)

(47)

(48)

(49)

(50)

(51)

(52)
(53)

(54)

107

J. L. Dye. M.T. Lok. F.J. Tehan. R.B. Coolen. N. Papadakis. J. M.
Ceraso and M. G. DeBacker. Ber. Buncenges Phys. Chem. 75.
3092 (1971).

J. L. Dye “Electrons in Fluids”: (J. Jortner. N. R. Kestner. Eds
Springer-Verlag. Berlin. 1973).

J. L. Dye. Science. 247. 663 (1990).

M. T. Lok. F. J. Tehan. and J. L. Dye. J. Phys. Chem. 76. 2975
(1972).

K. Tsai and J. L.Dye. J. Am. Chem. Soc..113. 1650 (1991).

C. D. Wagner. W. M. Riggs. L. E. Davis. J. F. Moulder. and G. E.
Muilenberg. “Handbook of X-ray Photoelectron Spectroscopy“:
Perkin-Elmer: (1979): p154.

R. D. Rieke. Science (Washington. DC.) 246. 1260 (1989) and
references therein.

H. Abe. K. P. Charle. B. Tesche. and W. Schulze. Chem. Phys. 68.
137 (1982).

R. H. Doremus. J. Chem. Phys. 40. 2389 (1964).

C. D. Wagner. W. M. Riggs. L. E. Davis. J. F. Moulder. and G. E.
Muilenberg. “Hand book of X-ray photoelectron spectroscopy”.
Perkin-Elmer.(l979). p. 154

J. C. Fuggle. E. Kallne. L. M. Watson. and D. J. Fabian. Phys. Rev..
816. 750 (1977).

C. H. Johansson and J. O. Linde. Ann. Phys/k. 25. 1-48 (1936).
S. Bhatia “Zeolite Catalysis: Principles and Applications“. CRC
Press. Inc. (1990)

D.W. Breck. W. G. Eversole. R. M. Milton. T. B. Reed. and T. L.
Thomas. J. Am. Chem. Soc.. 78. 5963 (1956).

 

(55)

(56)

(57)

(58)
(59)

(60)
(61)

(62)
(63)

(64)

(65)
(66)

108

a. l. E.Maxwell. Adv. Catal.. 31. 1 (1982).

b. P. Gallezot. Catal. Rev. Sci. Eng. 17. 273 (1978).

D. W. Breck. “Zeolite Molecular Sieves": Wiley: New York.
(1974).

G. Kumar. J. R. Blackburn. W. E. Moddeman. R. G. Albridge. and
M. M. Jones. Inorg. Chem. 11.296 (1972).

V. N. Romannikov and K. G. Ione. J. Catal.. 66.121 (1980).

J. C. Vedrine. M. Dufaux. C. Naccache. and B. J. Imelik. Chem. Soc.
Faraday Trans. 174.440 (1978).

J. H. Scofield. J. Electron Spectrosc.. 8. 129 (1976).

L. Broussard and D. P. Shoemaker. J. Am Chem. Soc.. 82. 1041
(1960).

N. S. McIntyre and M. C. Cook. Anal. Chem. 47. 2208 (1975).
A. Renouprez. C. Hoang-Van and P. A. Companon. J. Catal.. 34.
411 (1974).

G. H. Via. J. H. Sinfelt and F. W. Lyttle. J. Cem. Phys. 71. 690
(1979).

S. J. Gurman. J. Mat. Sci.. 17. 1541 (1982).

T.Sugimoto..Adv. Colloid Interface Sci. . 28. 65 (1987).

 

 

11111111111111