Note: Descriptions are shown in the official language in which they were submitted.
33~7~
This invention relates to par~iculate abrasiv~
agglomerates in which abrasive grits are held in a friable
matrix, ~uch agglomPrates being particularly suitable for use
in co~tad abras~ve products in which the agglomerates are
5 bondad to a flexible sheet backing; they can also be used in
bonded abracives (grindlng wheels~.
The use o~ small, particulate, agglomerates of
rela~ively ~ine abxasive grits held ln a matrix, for use as a
substitute ~or conventional abrasive grits on a coated
abrasi~e ("~andpaper") flexible abrasive,~was suggested at
least as early as U.S. Patent 2,194,472~ So far as is known,
the solid agglomerates of the type disclosed in the above
patent or products made from them have never been commercially
successful, U.S. Pa~ent Re. 29,808, discloses hollow spheres
(or o~her shapes, such as cylinders) consisting of abrasive
grits bonded onto the outer surface of a friable matrix, such
as resin or an inorganic ~ilicateO European published
application 8868, published March 19, 1980, discloses solid
agglomerates bonded by fused cryolite or other "salts or
silicates",
British Patent 982,215 and U.S. Pakent 3,156,545,
teach maklng ~ol~d agglomerates fox use in grinding wheels
consisting of glass bonded alumina or other grits. Benner
U.S. Patent 2,216,728 discloses g~ass or metal bonds for the
matrix of a~gregates containing diamond abrasive particles.
The patent states that ~he matrix may be made somewhat porous
to enhance mechanical bonding when the aggregates are mixed
with a binder to form a grinding wheel. U.S. Patent 2,806,772
~,
3~7
suggests includ~ng foamed glass in abrasi~e agglomerates
bonded by a resin matr~x.
W~ile abrasi~e agglomerates of the hollow resin or
silicate bonded type have shown good results in coated
abras~e applications, and agglomerates such as tauyht in the
~ e ~æc applicatlon show good results, both types of
agylomerates are difficult or expensi~e to manufacture and it
is dif~icult to control their friability.
More control of the physical properties of abrasive
agglomerates, and excellent grinding results in coated
abrasives can be achieved by providing agglomexates of
abrasive particles bonded by a foamed glass in which the
abraslve particles are contained within the walls of the
cellular glass ma~rlx. Such agglomerates can be manufactur~d
by mixlng approprlate abrasive grits with conventional known
compositions which produce a foamed glass structure upon
firlng. The glass composition, foaming agent, and, if
deslred, grindlng aid, are mixed together, formed into small
agglomerates ~ the desired shape, fired, and cooled. The
2a agglomerates may then be screened to appropriate sizes and
employed in a conventional manner to produce coated abrasive
discs, belts, or sheets. They may also be used to produce
resin bonded grinding wheels.
The present invention utilizes the basic friability
o~ the ~ellular glass and its controlled variability of
friability as a matrix for abrasive grit. When a cellular
glass is at the appropriate foaming temperature, it expands
and will stick t~ most materials around it. In addition, it
tends to encapsulate particles in its path~
This latter tendency is utilized when sized
abrasive grlt is mixed with a cellular glass batch and the
body brought to a cellulating temperature. Surprisingly, the
grit particles are readily distribu~ed throughout the cell
yet totally encapsulated by the glass in the walls.
Accordlngly, mixtures of various cellular glass
batches are blended with various volume percentages of grit,
the blended batch is pelletized to appropriately sized green
spheroids and those spheroids dried and fired to yield the
abraslve aggregate.
Cellular glass is sold as a soft abrasive in its
own right. Its major product qualities are its ready
~riability without catastrophic ~ailure such that upon
rubbing o~er a workpiece new sharp glass surfaces are
aonstantly being formed. In addition, the material is
lmpermeable so that there is no absorbtion of liquid into the
structure Abrasive agglomerate performance depends upon the
l.0 ~riability of the matrix~ Ideally, the matrix should
~racture or crumble as soon as the encapsulated grain begins
to lose i~s peak cut~ing quality. This invention provides
a product in which ~ine abrasive grit is encapsulated in the
foam cell walls as a discrete impurity. Ideally, ~he matrix
should be de~igned to exhibit a coefficient of thermal
exp~n~ion that is as close as possible to that of the
abrasive ~rit in order to minimize cooling flaws.
The subject grain can be formed into ex~ruded
chopped shapes, or can be formed into spheres. Friability
can ~e controlled by the ratio of pores to grain and/or the
ratlo of glass to grain. Higher density matrixes (60pcf~)
will tend to ~reak like a glass while lower densities will
increase friabillty.
While khe ~ize of the aggregates is subject to much
~ariation, dependlng on the particular application and grit
siæe, generally the aggrega~es will be 250 microns or larger
in diameter, since ~he foam glass process limits the minimum
s~ze glass-grit aggregate. The maximum size normally used
would not be over 5mm, at least in coated abrasive
applications. The abrasive grit will generally not be finer
than 1~ microns, nor coarser than 2 or 3 millimeters.
The preferred abrasive grit is fused aluminum
oxide, but co-fused alumina-zirconia alloy abrasives can be
used, as can silicon carbide abrasive grit.
As shown in tha example below, soda lime glass can
be used, but a non-devitri~ying alumino borosilicate
compos~tlon is superior~
3Z7
I'~e abraslve and glass mixture for forming the
agglomerates contains from 40 ~o 80 percent (dry bul~ volums)
of milled gla~s composition and from 20 to 60 percent of
abrasive grain. Up to 2Q percen~ addi~ion o~ a grindlng aid
such as cryolite can be added to such mixtures. The final
productt when in the form of spheres, will have a bulk
density of from 20 to 55 pounds per cu~ic foot (0.32 to
0 88g/cc).
The optlmum firing temperature and time depends
upon the particular composition us~d, the desired density
tPorosity) o~ the product. In general a temperature of 800
to 900 degrees C or higher for about 20 minutes is suitable.
The steps in a typical example of this invention
are as follows:
lS 1. Preparing a foamable glass batch by ball milling
soda lime glass cullet with 0.25% carbon black and 0O5% three
micron silicon carbide for 24 hours in a batch ball mill to a
median part~cle size of five microns or less.
2~ Adding a charge of 70~ by volume of the foamable
glass batch and 30% by volume of an abrasive grit, in
particular, a 180 grit fused dark aluminum oxide and blending
them dry at high speed. Subsequent to dry blending a 1~
addition of alum is added as a dilute liquid and wet mixed
followed by a 0.4% solids ad~ition of an aqueous
montmorlllonite slurry at a 4% solids content as a binder.
Sufficient additional water is added to palleti~e the mix to
a pellet particle size on the order of 20/40 mesh.
3. The generally ~phe~ical pellets thus formed are then
dried in a fluld bed dryer and dry mixed with an aluminum
3Q hydrate parting agent and fired in a rotary kiln at a
temperature of about 850 degrees C for 20 minu~es.
The resultant particles exhibit a specific gravity
of 30 to 35 pcf (.48 to .56 g/cc~, When examined
microscopically, it is observed that the glas~ tends to
encapsula~e the alumina particles in a foam bubble network.
It is also observed that the alumina particles tend to be
concentrated at the periphery of the bubble in a manner
3~
akin to froth floatation. The particles at the surEace are
still covered by a layer of glass.
The particles were screened to 20/30 and 30/40
UOS. sie~e ~raction then tested by using them as if they were
in themselves abrasive grits and maXing coated abrasive belts
in the conventional fashion. The bel~s were tested in a
standard metal finishing test system and compared with belts
made from l80 grit dark alumina.
It was found that the initial time to achieve a
comparable finish was longer for the aggr~gate belts than
grit belts but the total amount of metal removed and the belt
li~etime was between two and six times that of the grit belt
standard.
~epeated testing yie~ded erratic results, some
repeating the aforemen~ioned pexformance, others
substantially poorer. It was determined that the reason for
~the erra~ic performance was the tendency of the soda lime
~lass to de~itrify and the potential for the cristobalite
crystals to cause defects which sometimes caused the glass to
fail. Additional testing was made using the belt with an
aqueous lubricant and the resultant performance was
consistently bad. It was determined that this was caused by
the poor aqueous durability of the glass.
Accordingly, it was determined to use a batch that
would be essentially a nondevitrifying borosilicate made from
a mlxture of clay or volcanic ash and chemical additives
similar to that described in U.S. Patent 3793039. A mix o
66% volcanic ash, 15% kaolin clay, 5.5% 5 mole borax, 8%
dolomite, 2.7% lithium carbonate, 2% sodium bicarbonate and
l~4% carbon black was comilled. A l~ addition of liquid alum
was made prior to pellPtizing. The resultant pellets when
~ired at 330C, exhibited performance essentially similar to
those made from melted glass cullet which tested
reproducibly. In addition, when tested in a we~ enviroment,
the performance was reduced but still better than that of a
conventional belt made with 180 grit. The alumino
borosilicate has enhanc~1 aqueous durability.
3~7
--6--
It was fur~her found that a 10~ addition of
powdered cryolite enhanced the cutting perEormance.
Cryolite ls a well-known grinding aid for metals and is
apparently encapsulated in a fashion similar to that of the
alumlna grlt.
The following examples show that while not
pre~erred, slllcon carbide or co-fused alumina-zirconia
abrasive grlts can be used.
The ~irst experiment used the standard soda lime
glas~ foam mix to whlch we added 30~ 39 Crv~tolon 180 grit
and 10% fine cryolite. The product was foamed at about 850
degrees C. The resultant aggregate was lighter than that
made from alumina, its bulk density being 22.4 pcf @ 12/20
vs. 27-29 pcf @ 12/20 but seemed otherwise similar. It was
obser~ed that some of the grit particles were associated with
laxge bubbles which suggest that even coarse grit SiC will
influence the foaming reaction. These large bubbles will
probably weaken the aggregate.
The second experiment used the same ingxedients
except that 80 grit ofused alumina-zirconia containing 40
zirconia was used ~180 grit was not available). The
resultant aggregate was essentially equivalent to alumina in
all properties. A microscopic examination of the
encapsulated alumina zirconia grains showed that despite the
reduclng conditions of foaming, there was evidence of
oxidation of the metallic components of the grain. This
ef~ect wlll reduce the facture toughness of the grain.
Thus, it can be shown that other abrasives can be
encapsulated into aggregates as was alumina but there are
side e~ects that may reduce their utility.
Coated abrasive products are made from the
agglomerates of this invention by bonding the aggregates in a
single layer on a flexible backing sheet bv conventional
means well-known in the art, employing thermosetting maker
and si2e coats, glue, or a combination of glue and resin.
Subsequent ~o making the above examples, it was
found that silicon carbide containing aggregates, for certain
puxp~se~ ~uch a3 the grinding of titanium metal, are clearly
superior t~ conventlonal sillcon carbide coated abrasive
products.
Addit~onally it has been determined that when
glass forming chemical mixes, rather than pre-melted and
g~ound glass, are used in the form~ion of the aggregates,
superior wetting of the abrasive aan be achieved. In
addltion the resulting aggregate is different in structure
from the typlcal glass containing mixesO In the case of ~he
glass forming mixes, the abrasive particles are more
un~ormly dispersed within the multi-cellular aggregate body,
as compared to the glass mixes in which ~he abrasive
particles tend to be concentrated in the outer peripheral
cell walls of the multi-cellular matrix.
lS In a reduction to practice a foamable blend of
glass batch, at 70%, 180 grit green SiC at 30% was mixed dry.
To t~is mix 1% alum on a dry solids basis in aqueous solution
was added, followed by enough (0.4~) montmorillonite aqueous
~lurry to pelletize to a 20/40 mesh size. These generally
spher~cal particles were dried and fired ak 850Co for 20
mlnutes ~n a rotary kiln. If desired a lO to 20% addition of
cryolite as a grinding aid can be added at the dry mixing
stage.
The fired particles were coated on a belt in the
standard fashion and tested dry in finishing titanlum metal
and wet in ~inishing plate glass. In both cases there was a
longer break in period than that of a regular SiC belt but
the useful cut li~e was much longer. In the case of titanium
a standard belt cut 16 gm while the experimental nodule belt
cut a total of 245 gm. The wet cutting of glass was similar:
18 gm vs 180 gm for the experimental nodule beltO
The expeximent was repeated using a mixture of
chemlcals and minerals that yield an oxide glassy composition
known to achteve a good bond with SiC. This bond was ball
milled with ca~bon then blended, pelletized, and ~ired in a
slmilar ~ashion to the prior example~ The dry tests in
titan~um were equivalent to the glass matrix material but the
r
wet flnishing of glass was enhanc~d and a cut of 248 gm was
experisnced.
Microscopic observations showed the grit in the
glass to ha~e migrated substantially to the periphery of the
S bubble and the bubble center to be aomprised o a few large
Cell5, The hond composltion nodules had Sic distributed
throughout the nodule and the internal closed cells were
sm~ll and uniform in size.
