Note: Descriptions are shown in the official language in which they were submitted.
:~326~8
ABR~SIVE ~ATERIAL AND MET~OD
FIELD OF THE INVENTION
The invention relates to the production of
aluminous abrasive grits or shaped bodies containing high
density poly-crystalline alpha alu~ina, or such alumina --:
with other additives.
B~CKGROUND OF THE I~VENTION
Hard and strong abrasive ~rits, for use in
grinding wheels, flexi~le coated abrasive products
("sandpaper"), or as loose abrasive, are produced
commercially from alumina containing raw materials either
by fusion in an electric furnace or by the firing of
shaped bodies containing finely divided alumina at
temperatures well below the fusion points of the
material. Such lower temperature process is called
sinteringO Thi5 invention relat~s to aluminous abrasives ~-
made by the sintering process.
The first large-scale commercially produced
sintered abrasives were produced by the method taught in
U.S. Paten~ 3,079,243 to Ueltz. This patent teaches the
milling of calcined bauxite to produce a fine particle
qi~e raw material which is then formed into abrasive grit
sized particles and fired at about 1500C to ~orm hard,
strong, tough, pellets of poly-crystalline alumina. ~;
Recently, abrasive ~aterials consisting of grits
made up of alumina and magnesia spinel, presumably made
according to the teachings of U.S. Patent 4,314,827, and
made according to the teaching of ~ublished British
Application 2,0~9,012A, published l December 1982, have
30~ been commercially introduced. These materials are
produced by the sintering (at about 1400C) of dried
alumina gel particles. U.S. Patent 3,108,888 to Bugosh
aIso teaches making high density alumina (or alumina
containing) products by firing a dried alumina gel made
~from alpha alumina monohydrate (boehmite) or by hot
presslng dried powders made from such qels. ~
' : ':
~326~
-- 2
Alumina/magnesia-spinel commercial abrasives made
from gels contain alumina in the form of cells fro~ 5 to ~ -
15 microns in diameter. The cells, or "sunbursts", are
made up of many elongated alumina ar~s 0.2 to 0.4
micrometers in diameter (but some of which can be as large
as 1 micrometer in the form of very roughly spherical
"blobs"), the arms in each cell appearing generally to
radiate from the center of the cell. All of the arms in a
given cell are apparently crystallographically identically
oriented. Such orientation is shown by the fact all the
area of a given cell extinguishes simultaneously upon
sample rotation when viewed between crossed polarizers ~y -
transmitted light microscopy.
While the commercial a~rasives made from sintered
gels containing alumina and magnesia are high quality
abrasives, it has not been possible to produce high purity
alumina grits by the gel route. This is shown hy the
relative softness and lack of abrasive utility for the
"control" example 13 in U~S. Patent 4,314,827 which was
made from an alumina qel without metal oxide or metal salt
additions. -
The present invention is an improvement in the
art of making strong abrasive bodies whereby useful
abrasive products can be made fro~ alumina gel with or .:
without the addition of zirconia or spinel formers such as
magnesia~,
DISCLOSURE OF THE INVENTION
The invention resides in the discovery t~at
control of the micros~ructure of the fired product, such
3;0 ~that the celiular structure of the alumina in the prior
art abrasives is avoided, results in improved product
performance. The resulting product, instead of cell areas
of~5 to 10 micrometers in fliameter, contains alpha alumina
particles (crystallites) of submicron size (0.2 to 0.4 ,i
35 ~ micrometers). . -~
s~ ~ . ' ".
~ ~ .
. :, ::
:
8 ~ :
-. , ;.
In the case of the higher (e.g. in some cases at
5%) MgO additions these alumina particles are surrounded
by a matrix of spinel.
Conditioning of the gel to achieve the descrihed
effect can be achieved by vibratory milling of the mix in
the sol or diluted gel form while employing alumina bodies
as the grinding medium in the mill. It is believed that
the main effect of the milling is to introduce material
from the alumina grinding media into the alumina gel.
Also impurities such as zinc and iron are intro~uced from
the piping and associated equipment. Milling with -~
zirconia bodies, for example, is ineffective to produce
the desired essentially non-cellular structur~.
The first effective and reproducable method found
by us was to generate such material in the gel by
vibratory milling of the gel with alumina bodies. A
suitable vibra'ory mill is shown in U.S. Patent
3,100,088~ Typically the media may be 1/2 inch in
diameter and 1/2 to 3/4 inches long. The tub which
contains the media and the mix is vibrate~ in the
horizontal plane by an off-balance weight connected to the
shaft of a motor mounted coaxially with the tub which is
mounted on springs. The off-balance weight is mounted
a~acent the plane of the bottom of the tub end a second
weight is mountsd below it. The motor typically rotates
at 1200 rpm. The combined oscillation subjects the
contents to a milling action by the grinding media. The
interior surface of the mill is preferably lined, ~s with
rubber, to prevent contamination by erosion of the metal
30 ~ walls-
Various additives such as taught in U.S. Patent
; 4,314,827 and British Application 2,0g9,012A can be added
to the alumina before or after gelling. The most useful
additive presently known IS any compatible precursor of
: :,
~ 3 ~ 8
MgO, whereby the final product contains preferably around5~ MgO. The MgO is present, however, in the product as
spinel (magnesium aluminate: MgA1204) but calculated
as MgO in the analysis. Obviously lesser amounts of MgO
may be included since the alumina with no addition is an
excellent abrasive in its own right when produced
according to the present invention. The milled gel of the
present invention may serve as a matrix for various added
materials or abrasive particles.
The milled MiX may be simply poured or placed
into containers to dry and then broXen up into a~propriate
sized pieces by crushing, with undersized material
recycled to the beginning of the process. Alternatively,
the material may be formed or molded into shaped particles
as hy extrusion. In the case of extrusion, the rods
formed would later be cut or broken into appropriately
sized pieces. The minimum useful firing temperature is
significantly below 1200C, usually considered the
transformation point to convert to alpha alumina. The
upper limit is not critical so long as the fusion
temperature is not reached. Too long a firing or too high
a temperature can cause exce~sive crystal growth. Higher
temperatures increase the cost of the process also, so the
~ preferred firing range is 1200 to 1500C.
; 25 EXAMPLE I
In a large polymeric plastic mixing ~essel 30
pounds (13.6 Kg) of ~ondea SB Pural Alumina (supplied by
Condea) and 30 Imperial gallons (136 liters) of water were
mixedO This material was then gelled by adding 4.1 liters
of 14 weight ~ HN03. Magnesium nitrate hydrate (7.5
pounds, or 3.4 kg~ dissolved in 3 gallons (13.7 liters) of
water was then added to the alumina gel to give 5~ by
weight of MgO in the final product. It was mixed for 15
minutes and transferred to a Model M451 ~ weco mill and
~35 *
Trademar]c
:.
~ 3 ~ 8
-- 5
milled for 1 hour with 1700 pounds of alumina media. The
mix was recirculated through the mill for the one hour
milling time at a rate of about four gallons per minute.
After milling it was pumped into aluminum trays to a
thickness of about 3 inches (7.6 cm) for drying on
electric strip dryers.
The alumina media composition was about 90~ alpha
alumina with silica as the main impurity.
A series of batches according to the above
formulation were made up and combined for crushing and
firing.
The dried gel was then roll crushed and screened
to a through 14 mesh sizing before firing, to yield the
~esired final grit sizes. It was t~en prefired at 4~0C
for 16 hours and fired at 1400C for 30 minutes in a
rotary kiln.
After firing all of the products had a hardness
of 19 GPa (~ickers indenter, 500g load) and a very fine
microstructure in which there was no cellular
microstructure and, almost all of the alpha alumina was in
the form of generally equiaxe~ particles (crystallites),
0.2 to 0.4 microns in diameter, except Por rare square
blocky shapes about 5 microns in diameter. The blocky
~shapes may have indicated contamination. The product,
upon examination by the scanning electron microscope, was
seen to be comprised of a spinel matrix and a
~ .:
discontinuous phase of alpha alumina.
In some specific coated abrasive grinding
applications the material was superior to fused
30~ ~alumina~zirconia and superior to commercially available
sintered gel type abrasive of the alumina-spinel
compositionO
EXAMPL~
j Pural microcrystalline boehmite alumina, 22.7
~ ~
. ~ .
. .
. . - .
~, .. ,, ,1 ., ." ~., . j." . , .. , ! .,1, . : , . ..... ' . .. . . '
- 6 - 132 6~8~
~ .
kilograms, was mixed with 225 liters of water and 13.5
, liters of 14% HNO3 for 10 to 15 minutes.
i One half of the gel mixture was milled for 2
hours in the Sweco mill containing 1/2 x 1/2 inch ceramic
' 5 bonded alumina, 88 A12O3 (the main impurit~es being
¦ MgO 1.74~, sio2 8.9%, Fe2O3 0.18%, Tio2, 0.2%, CaO
0.8~, Na2O 0.34%), available from Coors Porcelain Co.,
and dried. This was the sa~e media as used in ~xample I.
The other half was simply dried without milling. The
10 dried gels were crushed to size, prefired at 450C for 16
hours and fired at 1400C for 1 hour.
The milled material had a hardness of 19.1 GPa,
the unmilled material had a hardness of ll.O GPa.
Material from each batch was screened to produce
50 grit abrasive grains which were then used to produce
vulcanized fiber ~acke~ coated a~rasive discs. The mille~
material outperformed commercial alumina zirconia abrasive
by better than 10% in grinding 1020 steel (the test showed
a 14~ higher metal removal). -
The un~illed product was inferior to fused
abrasive in all grinding tests, which was to be expected -
in view of its low hardness~
EXAMPLE III -
- - -
In an example similar to that of the milled
; 25 product of Example I, the gel was milled for 0.2 hours.
The pro~uct, fired at 1400C for one hour, was mainly of -
~ ;the fine random O.2 t-o- 0.3 mm crystal structure, but
; ~ ~ showed some cellular appearance,
.
~ EXAMPLES IV TO IX -
:-..:.
~ Further examples were performed in a study of the
effect of firing time at 1400C. All samples were made by
the general procedure of Example I. Condea
microcrystalline boehmite alumina was employed, milling --
:: ~ :: : .
was for two hours, but a~ter drying, the gels were
~ ~
. -: .
:. : .
- ~ :
~326~8~ -
7 ~
:.
prefired at 750C for 30 minutes. As the firing time was
lncreased, there began to appear in the product a coarse
lath shaped crystallization of alumina, randomly dispersed
among the fine 0.2 to 0.4 ~icrometer alumina particles.
The results are tabulated as follows:
Particle Size
Firing Time (Micrometers~Ratio
_(minutes)_ Coarse Fine~ Coarse/% Fine
l ~one 0.2-0.3 0
lO 3 1.0-2.0 0.2-0.3 5
~-5 0.2-0.320
4-8 0.2-0.350
Up to 8 0.2-0.380
~0 Up to ~ 0.2-0.395
Since the presence of the coarse fraction is
believed to be less desirable, the firing time at 1400C
should not be more than 5 minutes for the preferred
product when the material is pref;red at 750C for 30
minutes~
In all cases, no cellular structure was
observed. Th2 microstructures consisted of the
non-faceted submicron particles and the faceted lath-like ~`
~coarse crystals, excPpt in the ca~;e of the l minute firing ~-
where no laths were found. -i
.
By "non-aceted" we mean no regular faceting of
the crystallites was observed in a fractured surface at
~5,000X magnification~y the scanni~ng electron microscope.
The particles of alpha alumina were, instead, rather
formless, but equiaxed, with generally curved outlines,
30~ ~and very~few apparen~traight-~utlines. At 20,000X
magni~ication faceted structure begins to be clearly
apparent.
The abrasive grits o~ t~is invention have a
hardne~s measured ~y~the ~ickers indenter with a 500 gram
~ ~ ~
:
: : ,''':
~:
- 8 ~
load of at least 16 GPa (90% density) for alumina without
additions and at least 14 GPa for the grits which are
modified by the presence of 2~ or more of spinel formers
or are modified by other additives. I~ile pure dense
alpha alumina has a hardness of about 20-21 GPa, some
porosity may be desirable for certain applications which
would lower the hardness. When the alumina has a hardness
of 13 GPa or less, it is too porous for the purposes of
this invention. Preferred are hardnesses of 18 GPa or
higher as measured by the Vickers indenter ~t a 500g load.
EXAMPLE X
. . _ . _ :
A series of abrasives of varying magnesia content
were made.
The general procedures of Example I were
employed, including the milling (but for 2 hours) with
alumina media. In all cases the gels, after drying at
200C for about 30 hours, were crushed and screened and
then calcined at 450C for 16 hours. The resulting grit
sized particles were fired in a rotary kiln at 1400C.
The heat-up time to 1400C was about 15 rninutes, and the
time at 1400C was about 15 minutes.
Various amounts of magnesium nitrate were added
prior to the gelling. In one run no magnesium nitrate was
added. The MgO content and hardness of the abrasives were
as follows: `
Hardness
Run No. MgO Content % by Wgt. (Vickers 500 g load)
9498 0.14 19.9 ~ -
949g 2.50 19
9500 7.g5 19
, :. .
9502 12.71 19 -
In a series of tests of vitrified (glass bon~ed~ ~
grinding wheels employing 54 grit (a combination of 46 -
grit and 60 grit sizes) sized abrasive wheels made wit~
~35
.
:::,,
' : : :
~32668~
g
the above grits were compared with the highest quality
known fused alumina abrasive (sulfide process abrasive).
The tests were carried out by grinding slots in
tool steel (D3) at various controlled in feeds. In dry
grinding, at 0.5 mils (.0005 inches) downfeed, the
abrasive containing no added MgO (0.14~ MgO) had a
grinding ratio 16.18 times the grinding ratio of the fused
abrasive (G ratio is the volumetric ratio of material
removed to wheel wear). All of the MgO additions resulted
in superior performance over the fuised abrasive in the dry
grinding tests. In the wet grinding tests the
experim~ntal abrasives with MgO added were poorer than or
equal to the fused abrasive. At 2 mils the -~
no-magnesia-addition abrasive was superior to the fused.
In coated abrasive tests employing 50 grit size
abrasive (CAMI standard) an abrasive made according to
Example X, and containing 0.6~ MgO, incorporated into
flexihle abrasive discs performea better (136%) than
co-fused alumina zirconia abrasive on 1020 steel and
almost equivalent to fused alumina-zirconia on stainless
steel. The abrasives containing 2.5% MgO and 7.59% MgO
were also superior on 1020 steelO The higher MgO a~dition
was leis effective on stainless.
The 0.14~i MgO abrasive contained, in addition to
the alumina 0~25% Si02, 0.18% Fe203, 0.28% Ti02,
0.0S% CaO, and 0.04% Na20, presumably mainly introduced
in the milling operation. Similar levels of these
impurities were present in the other abrasives.
While applicants do not wish to be bound by any
particular theory of the invention, it is believed that
the introduction from the alu~ina media of particulate --
matter may effeat seeding of the crystallization of alpha
alumina during the firing. Additlonally, the other
impurities introduced in the milling step may inhibit
~35
` ~"
.::
.:
-:'
- lo- ~326~88
crystal growth of t~e final product by their presence at
grain boundaries between the alpha alumina particles.
As evidence of the fact that it is the debris
from the ~illing media ~hich is effective to condition the
gel so that it produces the desired high density, finely
crystalline, non-cellular alpha alumin~ upon firing at
about 1400C, additions of milled water were made to
alumina monohydrate together with acid, without milling of
*he gel.
Water, nitric acid, and microcrystalline boehmite
were mixed, as in Example lI, except that 6 batches were --
made, with varying additions of water containing the
debris worn from alumina grinding media, when milled for
several hours with water (no other addition to the water),
as follows:
"Milled water" additions to alumina monoh~arate
~Condea~:
Wt.Ratio of milling debris Wt. % Debris in Hardness
Trial to alumina monohydrate fired product* GPa
20 1. 0.0074 1.07 20+
2. 0,0037 0.53 20
3. 0.0019 0.27 19+
. 0.00075 0.11 17
0.00038 0.05 15 , ^
25 ~~ 6_ 0 0 12.S
*Note: Assuming an average loss of weight in
firing of 3~%~
The hardness was determined on the fired product,
fired at 1400C~+ 20C, for about 10 minutes. The furnace
30 i~was elect~rically fired, the atmosphere was air.
Examination of the milled debris showed it to be
;mostl~y~alpha alu~in~a with a surface area of about 39
square meters/gram.
H~igh purity~alumina produced by recovery of the
35~
,.-, :-
~.
~ 3 ~
11 --
fine suspended alumina particles left in suspension when
very fine alumina powders are permitted to settle after
being mixed with water is also effective, when used in an
amount of at least about 0.1~ of the fired gel solids.
Tests with commercial fine alpha alumina powders,
and tests with fine alumina generated by milling very high
purity fused alumina, using such alumina itself as a
milling medium, were very effective in producing the dense
finely crystalline product of the invention.
Differential thermal analysis has shown that,
when the alpha alumina seed particles are present the
transition of the gel alumina from presumably the gamma
form to the alpha form takes place at about 1090C, while,
in the absence of such seed material the transition takes
place at about 1190C. Thus, the theoretical minimum
firing temperature of the products of the present
invention can be below the usual reported transformation
temperature.
This invention, for the first time, permits the
manufacture by low temperature sintering of high purity
alpha alumina bodies having a submicron particle size and
a density of greater than 95%, resulting in a hardness
.
greater than 18 GPa. Products other than abrasive, such
as coatings, thin films, fibers, rods, or small shaped
parts, can be made by the process of the present invention.
Grain growth inhibitors such as SiO2 Cr203,
MgO, and ~rO2, have been added to the conditioned gel.
In ~he experiments in which MgO was added there was -
reaction with the alpha alumina and spinel was formed and ~-
~30 was observed as surrounding the remaining unreacted alpha
alumina. It was assumed that with the other additives
compound formation with the alpha alumina was minimal and -
they remained in the crystal boundaries. The experiments
which have been run clearly show that the crystal growth
,
. ~ ~ , . .
~ : ., .:
~ : :
' .
~L 3 ~
- 12 -
by diffusion and recrystallization was suppressed by the
additives. This has value in allowing more flexibility in
the time-temperature relationships for the sintered
products. The use of the growth inhibitors ~re well known
5 in ceramic technology and are not a necessary part of the
invention but are very useful to maintain the desired
microstructure over a broad range of sintering
time-temperature and where the high purity of alpha
alumina is not a requirement.
"'~' ~'
~
: .
^
: .
:: : '
' '.
