Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 02203170 1997-04-18
W ~ 96/12833 PCT11b5_/0~01
Stable Anodes for Aluminium Production Cells
_.
Field of the Invention
This lnvention relates to anodes for the electrowinning
of aluminium by the electrolysis of alumina in a molten
fluoride electrolyte, in particular cryolite.
The invention is more particularly concerned with the
production of anodes of aluminium production cells made of
composite materials by the micropyretic reaction of a
mi~ture of reactive powders, which reaction mixture when
ignited undergoes a micropyretic reaction to produce a net-
shaped reaction product.
Background Art
US Patent N 4,614,569 describes anodes for aluminium
electrowinning coated with a protective coating of cerium
oxyfluoride, formed in-situ in the cell or pre-applied, this
coating being maintained by the addition of cerium to the
_molten cryolite electrolyte.
US Patent N 4,948,676 describes a ceramic/metal
composite material for use as an anode for aluminium
electrowinning particularly when coated with a protective
cerium oxyfluoride based coating, comprising mixed oxides of
cerium and one or more of aluminium, nickel, iron and copper
in the form of a skeleton of interconnected ceramic oxide
grains interwoven with a metallic network of an alloy or an
intermetallic compound of cerium and one or more of
aluminium, nickel, iron and copper.
US Patent N 4,909,842 discloses the production of
dense, finely grained composite materials with ceramic and
metalllc phases by self-propagating high temperature
~ .
CA 02203170 1997-04-18
W O96/12833 PCT~B95/00801
synthesis (SHS) with the application of mechanical pressure
during or immediately after the SHS reaction.
US patent 5,217,583 describes the production of ceramic
or ceramic-metal electrodes for electrochemical processes,
in particular for aluminium electrowinning, by combustion
synthesis of particulate or fibrous reactants with
particulate or fibrous fillers and binders. The reactants
included aluminium usually with titanium and boron; the
binders included copper and aluminium; the fillers included
various oxides, nitrides, borides, carbides and silicides.
The described composites included copper/aluminium oxide-
titanium diboride etc.
PCT patent application W092/22682 describes an
improvement of the just mentioned production method with
specific fillers. The described reactants included an
aluminium nickel mixture, and the binder could be a metal
mixture including aluminium, nickel and up to 5 weight%
copper.
US patents 4,374,050 and 4,374,761 disclose anodes for
aluminium electrowinning composed of a family of metal
compounds including oxides. It is stated that the anodes
could be formed by oxidising a metal alloy substrate of
suitable composition. However, it has been found that
oxidised alloys do not produce a stable, protective oxide
film but corrode during electrolysis with spalling off of
the oxide. US patent 4,620,905 also discloses oxidised alloy
anodes.
US patents 4,454,015 and 4,678,760 disclose aluminium
production anodes made of a composite material which is an
interwoven network of a ceramic and a metal formed by
displacement reaction. These ceramic metal composites have
not been successful.
US patents 4,960,494 and 4,956,068 disclose aluminium
production anodes with an oxidized copper-nickel surface on
an alloy substrate with a protective barrier layer. However,
CA 02203170 1997-04-18
W O 96/12833 PCTAB9S/00801
full protection of the alloy substrate was difficult to
achieve.
US Patent 5,284,562 discloses alloy anodes made by
sintering powders of copper nickel and iron. However, these
sintered alloy anodes cannot resist electrochemical attack.
PCT application PCT/US~3/03605, as yet unpublished,
discloses aluminium production anodes comprising ordered
aluminide compounds of nickel, iron and titanium produced by
micropyretic reaction with a cerium-based colloidal carrier.
So far, all attempts to produce an electrode suitable
as anode for aluminium production and based on metals such
as nickel, aluminium, iron and copper or other metals have
proven to be unsuccessful in particular due to the problem
of poor adherence due partly to thermal mismatch between the
metals and the oxide formed prior to or during electrolysis.
Summary of the Invention
An object of the invention is to provide an anode for
aluminium production where the problem of poor adherence due
partly to thermal mismatch between a metal substrate and an
oxide coating formed from the metal components of the
substrate is resolved, the metal electrode being coated with
an oxide layer which remains stable during electrolysis and
protects the substrate from corrosion by the electrolyte.
The invention provides an anode for the production of
aluminium by the electrolysis of alumina in a molten
fluoride electrolyte, comprising a porous combustion
synthesis product deriving from particulate nickel,
aluminium and iron, or particulate nickel, aluminium, iron
and copper, optionally with small quantities of doping
elements, the porous combustion synthesis product containing
metallic and/or intermetallic phases, and an in-situ formed
composite oxide surface produced from the metallic and
intermetallic phases contained in the porous combustion
synthesis product by anodically polarizing the combustion
synthesis product in a molten fluoride electrolyte
CA 02203170 1997-04-18
W O 96/12833 PCT/lD53~0C~Cl
containing dissolved alumina. The in-situ formed composite
oxide surface comprises an iron-rich relatively dense outer
portion, and an aluminate-rich relatively porous inner
portion.
Comparative anodes of similar composition but prepared
from alloys not having a porous structure obtained by
combustion synthesis show poor performance. This is believed
to be a result of the mismatch in thermal expansion between
the oxide layer and the metallic substrate with the alloy
anodes. The differences in thermal expansion coefficients
allow cracks to form in the oxide layer, or the complete
removal of the oxide layer from the alloy, which induces
corrosion of the anode by penetration of the bath materials,
leading to short useful lifetimes.
In contrast, the porous anodes according to the
invention accommodate the thermal expansion, leaving the
dense protective oxide layer intact. Bath materials such as
cryolite which may penetrate the porous metal during
formation of the oxide layer become sealed off from the
electrolyte, and from the active outer surface of the anode
where electrolysis takes place, and do not lead to corrosion
but remain inert inside the electrochemically-inactive inner
part of the anode.
The composition of the combustion synthesis product is
important to produce formation of a dense composite oxide
surface comprising an iron-rich relatively dense outer
portion and an aluminate-rich relatively porous inner
portion by diffusion of the metals/oxides during the in-situ
production of the oxide surface.
The combustion synthesis product is preferably produced
= from particulate nickel, aluminium, iron and copper in the
amounts 50-90 wt% nickel, 3-20 wt% aluminium, 5--20 wt% iron
and 0-15 wt copper, and the particulate nickel may
advantageously have a larger particle size than the
particulate aluminium, iron and copper. Additive elements
such as chromium, manganese, titanium, molybdenum, cobalt,
CA 02203170 1997-04-18
W~96112833 PCT~B9S/00801
zirconium, niobium, tantalum, yttrium, cerium, oxygen, boron
and nitrogen can be included as "dopants" in a quantity of
up to 5 wt% in total. Usually, these additional elements
will not account for more that 2 wt% in total.
More preferably still, the combustion synthesls product
is produced from 60-80 wt% nickel, 3-lO wt% aluminium, 5-20
wt% iron and 5-15 wt% copper.
With the aluminium content in the preferred range 3-lO
wt%, the resulting composition has good adherence with
cerium oxyfluoride coatings when such coatings are used for
protection, and the lowest corrosion rate. Below 3%
aluminium, the composites still have low corrosion, but
surface spalling is found after testing. With increasing
aluminium content above lOwt%, corrosion increases
gradually, and above about 20wt% aluminium the composites
have low porosity due to the increase of combustion
temperature.
With below 5 wt% iron or no iron, the samples have
higher corrosion and a non-conducting layer is found after
testing. Above 20 wt% iron, results in surface spalling
after oxidation, 15 wt% being a preferred upper limit.
Below 5 wt% copper down to 0 wt% copper results in
anodes with higher corrosion rate but which are nevertheless
acceptable, and more than 15 wt%, in particular more than 20
wt% copper, results in surface spalling after oxidation.
When copper is present, it has been found that the composite
oxide layer is depleted in copper, whereas the unoxidised
portion of the combustion synthesis product adjacent to the
aluminate rich inner portion of the oxide surface is rich in
copper.
It is preferred to use very reactive iron and copper,
by selecting a small particle size of 44 micrometers or less
for these components.
CA 02203170 1997-04-18
W O 96tl2833 PCT/Lb5_t~C~01
- 6 --
It is recommended to use aluminium particles in the
size range 5 to 20 micrometers. Very large aluminium
particles (-100 mesh) tend to react incompletely. Very fine
aluminium particles, below 5 micrometers, tend to have a
strong oxidation before the micropyretic reaction, which may
result in corrosion when the finished product is used as
anode.
It is recommended to use nickel with a large particle
size, for example up to about 150 micrometers. Fine nickel
particles, smaller than 10 micrometers, tend to lead to very
fine NiAl, Ni3Al or NiOX particles which may increase
corrosion when the finished product is used as anode. Using
large nickel particles enhances the formation of Ni-Al-O,
Ni-Cu-Al-O, Ni-Al-Fe-O or Fe-Al-O phase on the surface which
inhibits corrosion, and also promotes a porous structure.
However, good results have also been obtained with nickel
particles in the range 10 to 20 micrometers; these small
nickel particles leading to a finer and more homogeneous
porous microstructure.
The powder mixture may be compacted by uniaxial
pressing or cold isostatic pressing (CIP), and the
micropyretic reaction may be ignited in air or under argon.
Excellent results have been obtained with combustion in air.
The powder mixture is preferably compacted dry. Liquid
binders may also be used.
The micropyretic reaction (also called self-propagating
high temperature synthesis or combustion synthesis) can be
initiated by applying local heat to one or more points of
the reaction body by a convenient heat source such as an
electric arc, electric spark, flame, welding electrode,
microwaves or laser to initiate a reaction which propagates
through the reaction body along a reaction front which may
be self-propagating or assisted by a heat source, as in a
furnace. Reaction may also be initiated by heating the
entire body to initiate reaction throughout the body in a
CA 02203170 1997-04-18
W O96/12833 PCTnB9S/00801
thermal explosion mode. The rea~tion atmosphere is not
critical, and reaction can take place in ambient conditions
without the application of pressure.
The combustion synthesis product has a porous structure
comprising at least two metallic and/or intermetallic
phases. Generally, the combustion synthesis product
comprises at least one intermetallic compound from the group
consiSting of nickel-iron, nickel-aluminium, aluminium-iron,
nickel-aluminium-copper and nickel-aluminium-iron-copper
containing intermetallic compounds
The porosity and micro-structure of the combustion
synthesis product are important for the in-situ formation of
the surface oxide layer since the pores accommodate for
thermal expansion, leaving the outer oxide layer intact
during electrolysis.
The porous combustion synthesis product may comprise
nickel aluminide in solid solution with copper, and
possibly also in solid solution with other metals and
oxides. Another material comprises a major amount of Ni3Al
2~ and minor amounts of NiAl, nickel, and a ternary nickel-
aluminium-copper intermetallic compound.
Other porous combustion synthesis products comprise at
least one intermetallic compound from the group A]Ni, AlNi3,
Al3Fe, AlFe3 as well as ternary or quaternary intermetallic
compounds derived therefrom, and solid solutions and
mixtures of at least one of said intermetallic compounds
with at least one of the metals nickel, aluminium, iron and
copper.
Another porous combustion synthesis product comprises
an intimate mixture of at least one intermetallic compound
of nickel-aluminium, at least one intermetallic compound of
nickel-aluminium-copper, copper oxide, and a solid solution
of at least two of the metals nickel, aluminium and copper.
.
CA 02203170 1997-04-18
W O 96/12833 PCT/lb
- 8 --
The porous combustion synthesis product may comprise
an intimate mixture of at least one intermetallic compound
of nickel-aluminium such as Ni3Al and Al3Ni, at least one
intermetallic compound of nickel-aluminium-copper such as
A173NilgCug, copper oxide, and a solid solution of two or
three metals nickel, aluminium and copper. It is believed
that the surface of this material and materials like it
contain non-stoichiometric conductive oxides wherein
lattice vacancies are occupied by the metals, providing an
outstanding conductivity while retaining the property of
ceramic oxides to resist oxidation.
Doping elements such as chromium, manganese, titanium,
molybdenum, cobalt, zirconium, niobium, tantalum, yttrium
and cerium may be present in solid solution or as
intermetallic compounds.
The in-situ formed composite oxide surface comprises an
iron-rich relatively dense outer portion, and an aluminate-
rich relatively porous inner portion which integrate into
the porous structure of the substrate. Analysis of specimens
~0 has shown that between the iron-rich outer portion and the
aluminate-rich inner portion is an aluminium-depleted
intermediate portion comprising predominantly oxides of
nickel and iron.
The outermost iron-rich oxide layer is a homogeneous,
dense layer usually comprising oxides of aluminium, iron and
nickel with predominant quantities of iron, preferably
mainly nickel ferrite doped with aluminium.
The aluminium-depleted intermediate oxide layer usually
= comprises oxides of nickel and iron, with nickel highly
predominant, for example iron-doped nickel oxide which
provides good electrical conductivity of the anode and good
resistance to dissolution during electrolysis.
The underneath aluminate-rich oxide layer is slightly
more porous that the two preceding oxide layers and is an
oxide of aluminium, iron and nickel, with aluminium highly
CA 02203170 1997-04-18
W~96/12833 PCTAB9~/00801
predominant. This aluminate rich layer may be a homogeneous
phase of aluminium oxide with iron and nickel in solid
solution, and usually comprises mainly iron nickel
aluminate.
_
The porous metal substrate close to the oxide layer
consists of nickel with small quantities of copper, iron and
aluminum. It is largely depleted in aluminium as the
aluminium is used to create the aluminate layer on top of
it, and is also depleted in iron. The metallic and
intermetallic core deeper inside the substrate is also
depleted of aluminium as a result of internal oxidation in
the open pores of the material and diffusion of the oxidised
alumlnlum .
The metallic and intermetallic core (deep down in the
sample) has a similar composition to the metallic core
nearer the oxide surface.
Interconnecting pores in the metal substrate may be
filled with cryolite by penetration during formation of the
oxide layer, but the penetrated material becomes sealed off
from the electrolyte by the dense oxide coating and does not
lead to corrosion inside the anode.
The invention also provides a method of manufacturing
an anode for the production of aluminium by the electrolysis
of alumina in a molten fluoride electrolyte, comprising
reacting a combustion synthesis reaction mixture of
particulate nickel, aluminium and iron or of particulate
nickel, aluminium, iron and copper (and optional doping
elements such as chromium, manganese, titanium, molybdenum,
J- cobalt, zirconium, niobium, tantalum, yttrium, cerium,
oxygen, boron and nitrogen) to produce a combustion
synthesis product which has a porous structure comprising
metallic and intermetallic phases, and then anodically
polarizing the combustion synthesis product in a molten
fluoride electrolyte containing dissolved alumina to produce
an in-situ formed composite oxide surface from the metallic
and intermetallic phases contained in the porous combustion
. .
CA 02203170 1997-04-18
W O96/12833 PCT/I~9~J'~01
-- 10 --
synthesis product, said in-situ formed composite oxide
surface comprising an iron-rich relatively dense outer
portion, and an aluminate-rich relatively porous inner
portion.
Another aspect of the invention is a method of
electrowinning aluminium by the electrolysis of alumina in a
molten fluoride electrolyte. The electrowinning method
comprises providing a starter anode which is a porous
combustion synthesis product comprising metallic and
intermetallic phases produced by reacting a combustion
synthesis reaction mixture of particulate nickel, aluminium
and iron or particulate nickel, aluminium, iron and copper,
and anodically polarizing it in a molten fluoride
electrolyte containing dissolved alumina to produce an in-
situ formed composite oxide surface comprising an iron-rich
relatively dense outer portion and an aluminate-rich
relatively porous inner portion.
Electrolysis of the same or a different molten fluoride
electrolyte containing dissolved alumina is then continued
to produce aluminium using the in-situ oxidised starter
anode. For example, the composite oxide surface is formed in
a cerium-free molten fluoride electrolyte containing
alumina, then cerium is added to deposit a cerium
oxyfluoride based protective coating.
In principle the final stage of production of the anode
will be performed in situ in the aluminium production cell
during production of aluminium. However, for special
applications, it is possible to form the in situ oxide layer
in a special electrolytic cell and then transfer the anode
to a production cell.
A coating may be applied to the in-situ formed oxide
layer; a preferred coating being in-situ formed cerium
oxyfluoride according to US Patent No 4,614,569. The cerium
oxyfluoride ~may optionally contain additives such as
compounds of tantalum, niobium, yttrium, praesodymium and
other rare earth elements; this coating being maintained by
the addition of cerium and possibly other elements to the
CA 02203170 1997-04-18
W 096/12833 PCT/lD5
1 1
molten cryolite electrolyte. Production of such a protective
coating in-situ leads to dense and homogeneous cerium
oxyfluoride.
Detailed Description
The invention will be further described in the
following examples.
F.x~rru?l e 1
A powder mixture was prepared from 73 wt% (68 atomic %)
nickel, - 100 mesh (<149 micrometer), 6 wt% (12 atomic %)
aluminium, -325 mesh (<42 micrometer), 11 wt% (11 atomic %)
iron, 10 micrometers particle size, and 10 wt% (9 atomic %)
copper, 5 - 10 micrometers particle size. After mixing, the
dry mixture (i.e. without any liquid binder) was uniaxially
pressed at a pressure of 176 MPa for a holding time of 3
1~ minutes.
The pressed samples were then ignited in a furnace at
900C to initiate a micropyretic reaction in air.
All reacted specimens were inhomogeneous and semi-
porous. Analysis of the specimens showed the following
20 ~ composition in atomic % : 59.8% nickel, 18.6% aluminium,
11.2% iron and 10.5% copper at the surface and 62.8% nickel,
13.9% aluminium, 12.3% iron, and 11.0% copper in the core.
The intermetallic compound NiAl3 was present.
Some specimens were then subjected to an oxidising
treatment in air at 1000C for several hours, typically 5
hours. Other specimens were not subjected to this oxidising
treatment, and it has been found that the oxidising
treatment is neither necessary nor preferred.
The specimens were then used as ~anodes in a cryolite-
based electrolyte containing 7 wt% alumina and 1 wt% ceriumfluoride at 980C. A typical test for a specimen with an
anode surface area of 22.4 cm2 ran for a first period of 48
hours at a current density of 0.3 A/cm2, followed by a
second period of 54 hours at a current density of 0.5 A/cm2.
CA 02203170 1997-04-18
W O96/12833 PCTAB95/00801
- 12 -
During the first period, the cell voltage was from 2.9 to
2.5 Volts, and during the second period the cell voltage was
from 3.3 to 4.4 Volts. At the end of the test, the anode
specimens were removed. The specimens showed no signs of
dimensional change, and the metallic substrate of dense
appearance was covered by a coarse, dense, uniform and well
adhering layer of cerium oxyfluoride.
After the electrolytic test, the specimens were
examined by scanning electron microscope and energy
dispersive spectroscopy (SEM/EDS).
The cerium oxyfluoride coating appeared homogeneous and
very dense, with no apparent porosity. On the surface of the
specimen, below the cerium oxyfluoride coating, there was an
in-situ formed complex oxide layer, total thickness about
300 micrometers, made up of three different oxide layers.
The outermost oxide layer was a homogeneous, dense
oxide-only layer devoid of fluoride. This~ oxide layer
comprised oxides of nickel, aluminium and iron with
predominant quantities of iron. The quantities of metals
present in atomic % were 32% nickel, 21% aluminium, 45% iron
and 2% copper. It is believed that this phase comprises
nickel ferrite doped with aluminium.
The intermediate oxide layer was composed of large
grains which interpenetrated with the outermost layer.
Analysis showed no detectable fluoride, and the intermediate
oxide layer comprised oxides of nickel and iron, with nickel
highly predominant. The quantities of metals present in
atomic % were 83% nickel, 3% aluminium, 13% iron and 1%
copper. It is believed that this phase is iron-doped nickel
oxide which would explain the good electrical conductivity
of the anode and its resistance to dissolution during
electrolysis.
The underneath oxide layer was slightly more porous
that the two preceding oxide layers. Analysis identified it
is an oxide of nickel, aluminium and iron with aluminium
CA 02203l70 l997-04-l8
W ~ 96/12833 PCT/l~
- 13 -
highly predominant. A small quantity of fluoride was
detected in the pores. The quantities of metals present in
atomic ~ were 22.6% nickel, 53.87% aluminium, 21.54% iron
and 1.99% copper. It is believed that this phase may be a
homogeneous phase of aluminium oxide with iron and nickel in
solid solution, forming an aluminate rich layer such as an
iron nickel aluminate.
The porous metal substrate in contact with the oxide
layer is comprised of nickel with small quantities of
copper, iron and aluminum. It is largely depleted in
aluminium as the aluminium is used to create the aluminate
layer on top of it. Its composition in atomic % was 77.8%
nickel, 5.3% aluminium, 3.5% iron and 13.5~ copper.
The metallic core deeper inside the substrate is also
1~ depleted of aluminium as a result of internal oxidation in
the open pores of the material and diffusion of the oxidised
aluminium. Here, the composition in atomic % was 77.2%
nickel, 1.8% aluminium, 9.7% iron and 11.3% copper.
All interconnecting pores in the metal substrate are
filled with cryolite, and in some cryolite-filled pores, a
second phase identified as aluminium fluoride is seen,
probably resulting from phase separation during the cooling
of the cryolite within the sample. No other metallic--
fluorides were detected in the metallic core.
2~ The metallic core (deep down in the sample) has a
similar composition to the metallic core nearer the oxide
surface.
Ex~m~le 2
The procedure of Example 1 was repeated varying the
proportions in the starting mixture, as shown in Table I.
The resulting specimens were subjected to electrolytic
testing as in Example 1. For the first five specimens, the
results were very good, and for the last two specimens, the
results were good.
CA 02203170 1997-04-18
W O96/12833 PCT~B95/00801
- 14 -
T~hle 1
r
Ni wt% Al wt% Fe wt% Cu wt% TEST
76.1 4.9 10 10
71.4 3.6 15 10
62 8 20 10 VERY GOOD
79 10 11 0
66.4 3.6 15 15
64 6 15 15
71 8 11 10 GOOD
Fxample 3
The procedure of Example 1 was repeated varying the
proportions in the starting mixture and with chromium as an
extra component. The particle size of the chromium was -325
mesh (<42 micrometer). The composition was nickel 73 wt%,
aluminium 6 wt%, iron 6 wt%, copper 10 wt% and chromium 5
wt%. Good results were obtained.
Co~D~r~tive Fx~Dle
Anode samples were made from nickel-aluminium-iron-
copper alloys prepared by arc-welding in argon. The
specimens were dense, non-porous and had the following
compositions in atomic % : 58.75% nickel, 23.17% aluminium,
9.19% iron, 8.94% copper; and 61.70% nickel, 14.86%
aluminium, 11.69% iron, 10.7% copper. Each sample was
oxidised for 5 hours in air.
The two samples were then tested as anodes in the same
conditions as in Example 1 at a current density of 0.3 A/cm2
for a period of 30 hours and 17 hours, respectively.
Both anodes were badly corroded at the =end of their
testing period. The reason the anodes did not perform well
during testing is probably a result of the mismatch in
CA 02203170 1997-04-18
W~96/12833 PCT~B95/00801
- 15 -
thermal expansion between the oxide layer and the metallic
substrate. These di~ferences in thermal expansion
coefficients allow cracks to form in the oxide layer, or the
complete removal of the oxide layer, which induces corrosion
S of the anode by penetration of the bath materials.
The porous anodes according to the invention, however,
accommodate the thermal expansion, leaving the protective
oxide layer intact, forming a barrier to further penetration
by the bath components. Bath materials which penetrate the
porous metal during formation of the oxide layer become
sealed off from the electrolyte and do not lead to
corrosion.
