CA1052984A - Metal oxide power densification process - Google Patents

Abstract

ABSTRACT
A powder, for example a mixture of titanium carbide and alumina powder, is densified into a fine grained sintered ceramic by applying a first constant physical pressure to the powder and heating it at a first predetermined temperature rate to onset of powder shrinkage, and then densifying the powder through at least two sequentially decreasing densification rates by variation of the applied pressure.

Description

105;~84 BACKGROUND OF TH~: INVENT ION
FIELD OF THE INVE~TION
This invention relates to materials and manufacturing pro-cesses for these materials and, more particularly, to an Lmproved uniformly fine-grain alumina-titanium carbide material and a ; technique for producing this material, and the like.
DESCPcIPTION OF THE PRIOR ART
Alumina (A1203) and alumina compounds have been used for . ~ .
high temperature and high strength purposes for many years. For example, in refrac~ory applications and in metalworking tools that are subjected to high speeds and great wear, these materials have found widespread industrial acceptance.
`~ It appears, moreover, that the strength of ~his material is in some manner related to its density and crystal size, the more dense and smaller crystal structure~ providing stronger and more durable tools. aonsequently, there is a great deal of emphasis on producing ceramic cutting tools with these characteristics.
When used as a cutting`edge, however, alumina occasionally fractures.
In general, these fractures seem to be related to the presence of relatively large alumina crystals, or "grains", in an essentially 6mall crystal or-"fine" grain structure. Thus, much of the alumina research efort has been directed to the more specific development of techniques for large-scale production of a high density material with a uniformly fine grain structure. ~ -The crystal growth that occurs when the raw powder mate~ial is heated to coalesce (or is "sintered") often i6 retarded through the addition of magnesium oxide (MgO) in an amount of 0.5% or less. This heating can be accomplished in a vacuum furnace that raises the material temperature to a 1400 to 1550C.

~ ~05'~984 ra~geO Processe~ o~ this sort have been reported to provide a material that has a crystal size on the order of 2 to 3 microns.
To attain this result, however, heating times in excess of four hours during sintering are required.
In the interest of efficiency and production economy, it is clear that a reduction in heating time is desirable, especially if the reduced heating time can be coupled with the pxoduction of a more unifonmly fine grain strueture. Because of the tendency .;. .
for alumina tools to fracture, there also is a need for a tech-i nique to produce the even smaller crystal size~ that ~eaa to ` greater strength.
~; SUMMARY OP T B INVENTION
In accordance with the invention, reduced heatang time and a fine crystal structure of æignificantly improved uniformity in size than that which heretofore has been available i8 achieved through a novel control of the physical pressure that i# applied . - .- ~ .
to powder to be sintered and the rate at which the pressurized - `
powder i# heated. Some materlal produced through tbis technique has c~mpres~ive and modulus o rupture strength3 that are signi-.
~t ` 20 ficantly greater than the best available alumina. - -;~ The proce~s characterlzing the invention i8 ~ essentially~
3~ a form of rate-controlled sintering in which a relatively low -pressure is applied to the die while the contained powder is ~ being heated. In the course of this heating the compacted powder ;3 ~ at first expands in volume. There is a point, however, tenmed ff~ the "onset o shrinkage temperature" or "onset o~ powder sbrinkage"
also called the "breaX away point", at which sintering commences and the volume of the powder begins to shrink. A maximum hot proce~s pres~ure is applied to the powder when this condition , . - . . ~, .
' .- . . ' , ~ " ~ : .

105'~984 is reached. Subsequently, the powder temperature also i8 increased to reach the maximum temperature attained in the process.
Thus, it seems that the physical pressure applied to the sintering powder lends an additional driving force that not only reduces production time, but also provides a demonstrably superior product.
In accordance with an embodiment of the invention, a pro-cess for densifying a metal oxide powder into a fine ground sintered ceramic comprises applying a compressing force to the powder, heating the powder at a predetermined temperature rate durin~ said force ~ , .
`', :':

~, .:

_ ~ 05'~9~4 application to an onset of powder shrinkage, and subsequently densifying the powder through at least two sequential decreasing den~ification rates by variation of said force.
In accordance with a further embodiment, a process for forming a fine grained sintered ceramic from a metal oxide powder comprises the steps of heating the powder at a temperature rate of 400 to 1000C per minute, applying a fonStant physical pressure to the powder while the powder is being heated at said rate until the powder reaches an onset of'powder shrinkage, subsequently sintering the powder through at least two sequential sintering rates and controlling the rate of sintering by adjustment of the applied pressure. -BRIEF DESCRIPTION OF THE DRAWINGS
Fig. 1 is a schematic graph of ram displacement versu~
time illustrate the "break away point" and Fig. 2 is an array of graphs that show pressure, tempera- -ture, density and breakaway point as function of time for a number of materials.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
- 20 Fig. 1 graphically illustrates features of the invention expressed in terms of the movement or displacement of the ram that i8 used to compress the powdered material which i9 being sintered as a function of time. The ram displacement 10 is necessary to "prepress" the powdered mixture in order to enhance sintering and to remove any entrapped gas'es in the powder between time 0 and tl. After time tl and before t~me t2, the application of heat to the precompressed powder leads to a thermal expansion displacement 12 of the ram. This step in the process is terminated .

lOSZ984 by a "break away point" 13 at the time t2. This "brea~ away point" is characterized by a change from the expansion of the precompressed powder to a contraction 14 that commences as sintering beginsO The contraction culminates at time t3. The time t3 is .a time of maximum densification and coalescence of the sintered powderO The further application of heat after ~ime ~3 produces excessive grain growth or "~loatin~" 16 as indi-cated by the increase in ram displacement. It is at this time : t3, before the material starts bloating, that the process is-~0 tenminated~ . .
Fig. 2 is a graphic representation of sintered product temperature, density and "break away points~ as a function of time ~or the following materials: -Billet Diameter Mate~xial . 1/4" U2 : lu A12 5" A1203 1" A1203-Tic . ~- ;

5" A123~
~0 E`or purposas of orient~tion between Figs.l and 2 the initial time, zero, of Fig. 2 corresponds to the time tl in Fig. 1.
~! The pressure "history" 20 for all of these materialæ is bounded ~ straight line segments that identify a pressure in-crease, a step ~unction from the initial pressure to the maximum hot process pressure that i8 maint~ined throughout the remainder of the process.
:, The temperature history 22 is bounded by straight line segments. These temperature bounds indicate an increasing tem-pera~ure in response to the initial heating, followed by a 105Z9~4 minLmum and maxLmum process temperature range for the remainder of the proces~.
The theoretical maxLmum density "history" 24 follow paths to maximum values which are represented by a generalized graph 24. The theoretical maximum density is defined as the closest possible packing of atoms into the crystalline structure of the compound, exclusive of any and all Lmpurities, that will pro- -duce a minimwm interstitial volume between the packed atoms.
.
s The brea~ away points as a unction time 30 vary, moreover ~
with the m~terial and billet size under consideration. `- ;
Exam~les:
Alpha alumina powder of less than one micron, preferably less than one tenth micron, particle size is worked or ball milled ~d in a dry mill fro~ four to eight hours. Preferably, alumina sold -~
by W. R. Grace Company under the name "Grace-KA 210" (trade mark) should be used as a raw material for the practice of the invention.
This alumina powder has a surface area on the order of 9 meters2~ -gram. It is, moreover, of very high purity, althougb it doe~
contain 0.1% addition of MgO. Other aluminas also can be used~
., ~ 20 although experimental data does seem to indicate that best re-.~ . .
sults are achieved with the Grace-KA 210 (trade mark) material.
~ To maintain powder purity, moreover, the ball mill also ;~ Yhould be foxmed from very pure alumina.
Upon completion of the milling step, the powder i8 baked `~
~ for another four to eight hours at 50 to 100C. ~akin~ the v~ powder at 72C. ~eems to be a preferred temperature for this 1~ step in the processO These ball milling and drying operations appear to have the effect of removing excess surface gases to produce a finer-grained end product. The relation between ' .

- ~05'~9~4 the surface gas and the yrain size of the fully processed material has not been definitely established. It is possible, however, ; that the surface gas behaves as an Lmpurity phase that causes severe selective grain growth at hi~h temperatures.
After outgassing, to produce a one-inch diameter billet of A12O3 in accordance with the invention, the powder is screened through a 200 mesh United States Standard sieve to break up any ; agglomerates that may have formed. The sifted powder is placed in a high temperature, high stren~th die. Typically, a graphite die in an inert, vacuum or reducing atmosphere i9 suitable for the purpose. A compacting pressure of 4000 to 8000 pounds per square inch (psi) is applied to the powder within the die. This pressure is applied to initially compact the powder to 30% to 50% of its maxLmum theoretical density. For thissample, it has been found that an initial compacting or Hpre-pressing" pressure of 5750 psi leads to the best end product results. This prepress force is then reduced to a range of 500 to 1000 psi. Generally, a reduction in pressure to 1000 psi will produce acceptable -~
results.
~he powder and the die are placed in a hot press or other high temperature and high pressure sintering device. A protective atmosphere, moreover, is established in this system in order to preserve the die. A vacuum, a helium or other inert atmosphere, or a mixed atmosphere of inert gas and 8% by weight of hydrogen have been found suitable for this purpose. Furthermore, relative-ly less expensive nitrogen gas may be used for process economyO
Starting then with the reduced pressure on the compacted E
powder, the temperature of the powder and die is raised by means of an induction heater at a rate that is bounded by 400 to 1000 - 105'~4 per minuteO By rrGper po~itioni~ and sizing of the induction heater and the billet generally uniform heating throughout the powder can be established. Within the above range it appears that the rate of temperature chanqe can be varied in an almo~t random manner until the onset of shrinkage of Ubreak away point" ' 13 (~igO 1) is reached without degrading the quality of the final product.
` With respect to the sample under consideration,numerous tests indicate that raising the temperature, within the above `' 10 rate boundaries, of the powder and the die to 760 to 815 C. as ;~
measured with an optical pyrometer will pxoduce the desired re~
. . .
sult.'That is, the onset of shxinkage or "break away point" usually commences as the temperature reaches about 800C. In accordance with a feature of the invention, while the temperature is being raised to the illustrative 800C., to commence shrinkage, the reduced pressure of 1000 psi also is applied to the powder billet.
This shrin~age may bo observed with the aid of a linear variable displacement transducer that is attached to the ram that applies ' ~' the pres6ure-to the sintering powder.
After the ~break away point" is reached, both temperature and pressuxe are increased in order to promote the rate of den- ' sification that is inherent or natural to the particular material and billet size. Both pressure and temperature can be monitored and adjusted to approximate this natural rate.' ~his natural den-sification rate is identified through a serie~ of tests conducted with sample powders. In each of these tests, pressure and temperature increase rates are varied to identify the ranges o~
pre~sures 20 (Fig. 2) and temperatures 22 that provide the closest approach to the theoretical maximum density 24. It should .
.

- 105'~984 be noted in FigO 1 that the natural rate of den~ification c~anges as the powder is sintered into its maxinum densification as indi-cated by the minimwm billet volume at time t3.
With respect to the above alumina example, the onset of powder shrinkage is accompanied by an application to the n~w sintering billet of a physical or ram pressure of 3600 psio Al-though this a preferred maxLmum process pressure, suitable results are obtained with pressures in the range of 2000 to 6000 psi.
This rapid increase in pressure is reflected in the step-function pressure change that characterizes the pressure graph 20.
As the application of this pressure continues, the tem-perature also is increased, but at a lower rate than that which ~ cbaracterized the initial increase to 800C~ Best results seem .,~ .
to be achieved with a temperature of about 1600C. that i~ reached about eight minutes after the earlier 800Co temperature wa~
attained. ~hese higher temperatures also are observed through an optical pyrometer. This maximum temperature and pressure are sustained for two to six minutes, and preerably for three minutes, if a maximum process temperature of 1600C. is achieved.
During this time, the alumina is sintering at its "natyral" or inherent rate of densification.
The linear change in ram displacement between the times t2 and t3 ~hown in Fig. 1 is a characteristic feature of a billet that is sintering at this natural rate. Other natural densification rate indices are possible> although ram displacement is a most convenient technique.
In accordance with the invention, from a broad viewpoint the pressure and temperature that are applied ta the sintering billet after the "brea~ awa~ point" 13 has been reached are _ g _ ' ` ' i ... .

` 105Z984 adjusted to establish and maintain this natural densification rateO The natural densification rate will, of course, vary according to the material that is being processed. This natural rate, moreover, also may vary for different batches of the same materialO Consequently, the precise temperature and pressures r that should be applied to the sintering billet for any particular material can be deter~ined through a number of tests each per-formed on a different batch of the material. These test~ will identify those conditions that produce the linear ram displace~
ment 14 (Fig. 1), or other indications of the naturàl den~ifi-cation rate, for the material under consideration. Once these sintering conditions axe identified, cubcequent billets can be processed without ram displacement observations and the like.
Thus, the natural rate of densification can be approximate-ly identified by applying a first physical pres~ure and heating ~j rate to a first batch of the material to be~in the intering of said first batch, applying an increased second physical pressure ~ -~ . . :. ;
and a second heating rate to said first batch, recording the ~-- . .
density of the sintered first batch, applying a physical pressure and heating rate to a second batch of the material to begin the sintering of said second batch, and applying an increased third physical pressure and a third heating rate to said second batch, recording the density of said sintered second batch to identify the density of said first and second batches of the material that is closest to the theoretical max~mum density of said materialO
A more detailed consideration of Fig. 1 indicates that the ram displacement is not entirely linear toward the completion ~;
of sintering 17. Thus, as shown in the drawing, the rate of ram displacement as a function of time decreases a9 the 9 intering ~ ~OSA~984 billet approaches a condition of maxLmum densificationO As this terminal portion of the sintering process is approached, the pressure and temperature applied to the billet i5 stabilized for two to six minute~ to "cure" the now sintered billet.
Thus, in the process of Fig. 1 to densify a metal oxide powder into a fine grained sintered ceramic, during stage 12, a compressing force is applied to the powder and the powder is heated at a predetexmined temperature rate during said force application to an onset of powder shrinkage at 13, and the powder - 10 îs subsequently densified (sintered) through at least two sequential decreasing densification (sintering) rates, during the sintering and curing stage 14, by variation of said force. The densification rate is constant, or linear, until the terminal portion at 17 is approached when the densification xate is non-lineax, the rate of sintering being controlled by adjustment of ~ the applied pressure.
J` Care must be exercised to terminate production conditions ; at this point in order to prevent the development of a "bloated"
billet. This "bloating" 16 is characterized by a reduced density . ~
billet, as indicated through the greater billet volume which the increa~ing ram di~placement registers. ~-~
Turning once more to the completion of sintering 17, it is possible to more precisely promote the natural ratc of sinter~
ing, which apparently changes as maximum densification is ap-proached, by adjusting the temperature and pressure that is applied to the sintering billet in a manner that will enable the ram displacement to more nearly approximate the preferred curve illustrated in Fig. 1.
After the period of curing, or sustained heating at the 105'2984 maxLmum process temperature and pres~ure, the induction heater, or other source of heat, is turned off and the pressure on the, alumina wi~hin the die is reduced to zero. A cooling period of one to five minutes is sufficient to enable the die (and the now sintered alumina) to cool to room temperature for removal from ,,the press and separation from the die.
'ASamples of sintered alumina, produced in the foregoing ,~'manner, have shown in carefully executed laboratory test~ the '~ following characteristics: -~ 10 Nwmber'ofAvg. KnPllOO*
','~ ' SamplesHardne~s . . .
Grace-KA 100 (trade mark) 8 2045 ~- Gxace-KA 210 (trade mark) ' 21 2334 ~
Commercial Sample A 10 2277 ' ~, Commercial Sample B 10 1952 ~ -~', *' Xnoop hardness is a measure of the microhardne~s of a '~ material by means of a long, narrow~ diamond ~haped impression.
The hardness number i8 calculated as the ratio of the indenting load to the,projected area of the indentation: THE MAKING~
SHAPING AND TREATING OF STEEL, ~NITED STATES STEEL 8th EDITION.
In thi~ connection it should be noted that the tenm "standard deviation" as used herein is the square root of the arithmetic mean of the ~quares of the deviations o~ the physical test data from their arithmetic mean.

-- ,- - - ~ : :

' ~ .

AvgO Compressive Avg. Modulus of Strenqth psi . Rupture. ~si Grace-KA 100 (trade mark) 326,700 44,100 Grace-K~ 210 (trade mark) 543,20~ 82,600 . Commercial Sample A321,000 59,500 : Commercial Sample B404,300 65,700 Modulus of Ruptuxe Standard Deviation, p8 i.
Grace-KA 100 (trade mark) 1~,500 : Grace-KA 210 (trade mark) 23,200 Commercial Sample A 16,000 ~ ~-- Commercial Sample B 11,300 .~ .
Compressive Strength Standard Deviation, psi Grace-KA 100 (trade mark) 115,000 Grace-KA 210 (trade mark) 122,300 -Commercial Sample A 111,600 .:
Commercial Sample B 104,200 -~ -Average Grain Size : :
~ . ~ ~

Grace-KA lOO (trade mark) 2.6 Grace-KA 210 (trade mark) 0.72 c . Commercial Sample A 1.3 Commercial Sample B 1.7 s The superior properties, on the average, of the sintered .. alumina that can be obtained if the Grace-KA 210 (trade mark) powder is used as a basic raw material in the procesa characteriz-ing the invention i9 apparent. It should be noted that the 105'~984 Grace-KA 100 (trade mark) powder does not have an added 0.1%
MgO crystal growth inhibitor. In developing the foregoing test data, moreover, sample preparation has been found to exert a signiicant influence. Chemical polishing of the samples, for instance, provides more realistic modulus of rupture test data.
Mechanical polishing, however, seems to be detrimental to the actual strength o~ the sample that is undergoing testing.
Studies with a scanning electron microscope (at a magnifica-~tion of 10,000) of the fracture surfaces of representative samples of alumina ceramic billets in the 1" to 5" diameter range that were produced in the manner described above demonstrate that the material has a grain size distribution as follows: -Grain Size Ran~ePercent of Grain Structure Less than 0.3 micron 0%
Between 0.3 and 005 micron 25%
Between 0.3 and 0.7 micron 54%
Between 0.3 and O.g micron 80%
Be~ween 0.3 and 1.5 micron 100%
i The "break away point" graph 30 in Fig. 2 illustrates the '~ 20 relation between the diameter of the end product billet and the process conditions. Thus, to manufactuxe a five inch diameter billet of A1203 in accordance with the principles of the inren-tion, somewhat higher temperatures and pressures shouId be applied during processing than these conditions which are mentioned above with respect to the oneinch di~meter billet~ It should be kept in mind, however, that a ~asic feature of the invention for all of the materials and billet sizes described herein is the appli-cation of an increased process pressure, within described boun-daries throughout the sintering process, i~e~ after the Ybreak ... .

away point" (Fig. l)o ~oreover, a maximum process pressure, an observed optimwm, is identified within the described bounds ob-tained by comparing the pressure "history" of the sintering billet with the density of the billet, and may be more conveniently applied to the billet to provide the desired closest approach to the theoretical maximum density.
Thus, alumina ceramics manufactured in accordance with the principles of the invention have a grain structure that is different from those grain sizes that have characterized the prior art. Crystals of much larger average size, e.gO two or three microns, ordinarily were grown in these pxior art alumina.
Accordingly, a new alumina ceramic with a ine grain size and -~ better grain size distribution that hexetofore was unobtainable is provided through the invention.
~ he invention, moreover, is not limited in application to alumina but also can be used in connection with other metal-qxides.
For example, uranium dioxide (U02) pellet fabrication can be im-proved through the practice of the invention. Typically, a pellet density that is within 1/2% of the theoretical attainable maximum ` 20 can be reached by means of this pressure and temperature rate con-; trolled sintering. Illustratively, to achieve 95% of the theore-tical maximum density, the powder is subjected to maximum process temperatures that are on the order to 800 to 900C in an eight to nine minute heating cycleO Within this time cycle, moreover, physical pressure also is applied to the powder that is being sintered. ~here is, of course, an initial or preliminary heating - period of about one minute, characterized by the onset o~ powder shrinkage, during which time the powder is raised rapidly to a higher temperature and subjected to increasing physical or --` 105;~984 mechanical pressureO The re~ulting uranium dioxide pcllets do not require grinding or other finishing operations because they ~- are fabricated in dies of correct diameter~ The elImination of a machine finishing operation in the fabrication of uraniwm dioxide reactor fuel pellets is especially ~eneficial because it reduces processing costs and eliminates a major source of fissionable material manufacturing waste.
A further example of the invention comprises the sintering ` of alumina with other carbides, nitrides or oxides to improve further the physical properties of the resulting product. As a specific example five inch diameter billets of alumina-titanium :
carbide (A1203-TiC~ were made from 70% alumina powder Grace-KA 210, trade mark and 30% titanium carbide powdex. The original particle ~ize of the titanium carbide powder is 2 to 4 microns. The par-; ticle size is reduced by ball milling for 16 hours in alcohol, to an average particle size of 1 micron. The ball-milled powder is mechanically mixed with the alumina powder for uniform distribution of the two materials in the resulting powder. Illustratively, the alumina and the ball milled TiC are blended together in an alcohol mixture in a ball mill for four hours. These mixed materials are removed from the ball mill, the alcohol is evaporat-ed and the resultinq powder, having thus been worked to remove surface gases and to reduce agglomerates in the powder, is pre-pxessed or compacted with a pressure in the range of 4000 psi to 8000 psi to achieve a prepressed billet that has a density`that is 30% to 50% of the maximum theoretical density. For the example under consideration, the 6300 psi pre-pressing pressure affords a suitable balance between powder packing and the elimination of en-trapped gases. The applied ram pressure is then reduced to a range 105;~984 of 500 to 1000 psio While this lower pressure i8 being applied, the material is heated at a rate that is not less than 400C per minute nor more than 1000C per minute until the onset of shrinkage commences, usually at about 800Co While the material is being heated to this 800C temperature the aforementioned reduced pressure is maintained constant to provide billet integrity, as noted aboveO
With the onset of powder shrinkage that occurs at point 31 on the "break away' graph 30 in Fig. 2, the ram pressure on the now sintering billet is increased to 5000 psi, the preferred maximum hot process pessure. Suitable results, however, can be obtained with applied ram pressures in the 3000 to 9500 psi range.
As the application of this pressure continues, the temper-` ature is increased, but at a lower rate than that which charac-terized the initial increase to 800C. Thus, within six to ten minute~, the maximum process temperature is reached in the range , .~
from 1200C to 1800C. Based on available experimental dataJ
best results are achieved with a temperature of about 1500C. Fox curing this maximum temperature and 5000 psi pressure axe sustained for two to six minutes. Thus, the powder is sintered at a rate of densiication that approaches the theoretical maximum density of the powder until material densification is complete~
As shown in Table I below, the resulting material is sup-erior to chemically similar materials that are produced through :
prior art processes.
Twenty, five-inch diameter billets of alwmina-titanium ca~bide were fabricated in accordance with the principles of the invention to demonstrate process reproducibility and the superior physical characteristics of the product.
~ he resulting density dat for all 20 billets, is shown :, ~

105;~984 in Table I. ~he average billet density was 4.257 g/cc -0.07%, whereas the prior art density for this material is 4.21 g/cc.
The term average, as used herein, is the quotient of the arithmetic sum of the data divided by the number of data values used in ca lcu lat ing the s um O
TABI,E I . B ILLET DENS I~ IES
BILLET NUMBER - DENS IT IES ( q/cc ) .. .. . . .................................................... .
63 4.249 64 4.259 1~ 65 . 40254 , 66 4.257 67 ` 4.256 ' 68 4.260 69 4.260 4.256 : 71 . 4.25~ - .
72 . 4.250 73 4.258 74 `` 4.260 4.262 ~ 76 4.257 .~ 7'~ 4.258 ;' 78 4.258 4 ~ 256 4.260 81 4.254 . 82 4.260 Average 4.257 gm/cc S~andard Deviation 0.003 gm/cc : (0.07%) _ 18 --- 105;~984 The billets were ground top and bottom on a slanchard model No. 11 grindex and diced into 21 blank~, each 3/4" square and 5/16" thicko From each of the 20 billets, two of the 21 blanks wexe randomly selected for transverse rupture strength testsO (TRS)o The two selected rupture test blanks were each sliced into three 1/4" by 3/4" by 5/16" parallelepipeds to pro-vide a total of six rupture specimens for each billet. The specimens were surface ground on all sides for edge sharpness and size uniformity.
The individual specimens were tested for transverse rup-ture strength by a three-point loading. The transverse rupture strength (TRS) results of these tests are tabulated in Table II. In Table II, the average TRS of the 8ix rupture ~pecLmens .. . .
~ taken from each billet is tabulated below along with the standard . . . .
deviation for this dataO The overall average (the average o the average of each group of 8ix samples) and the standard de-viation of this overall average was found t~ be 124,333 +

11,542 ps i . TABLE II
BILLET TRANSVERSE RUPTURE STRE~GTHS
~, . . .
20 B~LLET N0. SAMPLE ~0. TRS (PSI~ AVE. T~S (PSI) ~` 63 1 114,379

2 145,899

3 89,388 124,557 -~

4 133,232 -18,750 139,993 6 124,453 ., .
64 1 139,002 2 150,663 - 19_ -- 105'~984 BILLET N0. SAMPLE N0, TRS (PSI) AVE:.TRS ~PSI) 3 146,399 l46,187 4 137,252 - 5,967 149,353 6 154,456 1 `136,167 2 150,363 3 120,494 -- 113,799 - 4 104, 129 ~28, 545 .. 5 59,86 - 6 111,783 66 1 89,962 2 127,815 ~.
3 78,107 117j929 - 4 132,966 +24,59 ~ 5 14~,334 `~ - 6 136,392 67 1 119,570 ., . . .
~ 2 124,848 :~l 3 87,309 97,015 `~ 20 4 70,860 +20,937 ;. 5 '9 -6 74,539 68 1 135,543 . 2 139,812 3 125,472 132,003 4 118,659 + 7,94 ' ' ' ' . :

-- 105;~984 BILLET N0. SAMPLE N0.TRS (PSI~ AVE.TRS (PSI) 131,558 6 140,972 , . 69 1 136,554 2 106,444 3 89,163 117,195 4 148,363 ~21,286 .:`,9 - :
~ 6 97~147 ., . :
7 1 161,221 2 147,558 .
:
, .
.~. 3 84,116 130,232 : .-~`~ . 4 108,6l4 +26,180 --;
~ 5 145,953 ;~ 6 133,870 :.
. 71 1 :141,234 2 . 128,750 ~- 3 1~5,508 137,381 -. 4 131,122 + 6,832 149,525 ,, ::: -.
6 138,146 ~9~ 72 1 .119,675 2 126,684 3 68,657 1o8,767 4 142,746 -27,271 -86,07~ ~
`:
3 6 No Test r , ' .- ' - 21 _ .

.. . . . . .

105;~984 , B ILLET N0 . St~M PLE N0 . TRS ( PS I ) AVE . TRS ( PS I ) :
73 1 144,735 2 146,156 3 132,574 13, ,011 4 151,390 +12,447 6 113,596 74 1 116,096 2 136,419 3 99,~3.5 120,030 4 118,341 ~14,520.
. 5 140, 984 - 6 108,405 - 75 1 149,731 .
~. 2 132,475 : j 3 143,868 143,487 .j 4 151,635 + 6,244 . 5 141,792 `. 6 141,422 :~

76 1 121,548 2 133,975 3 79,607 121,8Z4 `1 4 139,120 +19,871 ~ 5 123,104 .; 6 133,589 77 1 132,863 2 141, 150 . - 22 -~` 105;~984 B ILLET N0 ~ SAM PLE N0 . TRS ( PS I ) AVE . TRS ( PS I ) 3 107,048 130,709 4 111,855 +15,711

5 147,159

6 144,178 . 78 1 137,132 ` 2 118,460 3 113,052 120,846 ?
`` 4 108,541 +18,356 . 10 5 96,384 . ~ `-6 151,509 79 1 103,270 :
. 2 118,557 3 110,070 120,667 4 124,957 +11,877 139,715 ~ 6 1~7,435 . :-~

-? 80 169,338 : ~:
- 290,804 s 20 . 3152,226 119,563 - -~ 4110,484 +31,824 ; :- ~
.~ . .
'~ 5151,916 : -. - -. , .
6 142,611 . 81 1 140,372 ~ 2 141,461 ` ``
i 3 136,273 130,407 4 138,052 +16,908 :'- :- , : ' ' - -` : ~ - ,:

- 1~5;~984 B ILLET N0 o SAM PLE N0 ~ TRS ( PS I ) AVE: .TRS ( PS I ) ~;, 93,o83 6 133,200 82 1 141,398 2 78,340 3 139,642 117,042 4 88,848 +25,437 - 5 115,036 6 138,986 ~- - 10 The broken transverse rupture specimens were then mounted and polished for hardnes~ testing. RockwellA hardness tests were discontinued when three of the Rockwell indentors were ruined after application to only five billets. In passing, it should be noted that a RockweIl test is a measure of hardness as mani-ested by the materials resistance to the penetration of an ~-~
indentor in response to the application of a known load. The subscript~ A in this test, indicates the load and indentor type used in the test for this material (TEE MAKI~G~ SHAPIN~ AND
TREATING OF STEEL, UNITED STATES STEEL 8th EDITION, 1964). Knoop hardness test~, however, were performed on all twenty billet~
: . . ~., The hardness data are tabulated in Table III.
... . .Although the RockwellA testsare not conclusive due to the above mentioned breakage problem, the average of five data points indicates a 0.8 increase in the RockwellA hardness over the prior art. ml8 0o8 increase is a ~ignificant improvement over the ` -psior art because increase~ o~ Ool are of practical importance in the industry, e.g., tools are graded by increa~es of 0.1 in RockwellA hardness.
~4 ~
., .

~.
TABLE III. BILLET I~ARDNESS
B IL~ET_ NO . R . HARD~ESS KNOOP HARDNESS
` 63 93075 3557 64 93.78 3557 `; 67 93-95 3557 .,, 68 No D. 3557 .. 69 N. D. . 3557 ~ lo 70 N- D~ 3557 . . -71 N. D. . 3227 73 N. Di 3557 -~
74 N. D. 3557 ~ , . , . ., "
N. ~. 3557 ~-~, 76 N. D~ . 3557 ., . . - `-. .
` 7? N. D. 3227 . ;~
78 N. D. 3557 ` ~ ~ ~
. 79 N. D. 2940 . ~ ~
,~: . - . ' ,-:
: .20 80 N. D. 3227 , . ~,. ~ ,.. .
.~ 81 N D . 3557 .~
.j~ : 82 N. D. 3557 -~ Average 93.82 N. D. = Not Determined Average 3477 ~ `
Two of the six broken transverse rupture spec~mens ~from ~ ..
~i each billet were photographed at a ten power (lOX) magnification, ~ Sample A, for macro-homogoneity, i.e. visible differenceY in ;~ the color of the sample material under inspection. Only one specimen from all of the samples s.tudied showed an inhomogeneity 5 - ;

, .
~: . - . . .... . . . . . . .
: . . , ^~ 105'~9~4 (a 004 mm equivalent diameter titanium carbide particle) as enumerated in Table rv below. The equivalent ~ize of the in-homogeneities listed in Table IV, moreover, are defined as the average of the major and minor axis of the inhomogeneity.

TABLE IV. BILLET MACRO-HOMOGENEIlqr NU~BER OF VIS IBLE EQuIvALENT
:~ DIl~F~RENCES IN S IZE OF
BILLET NO. SAMPIE A ~O. COLOR INHOMOGE:NEITY
63 1 o __ lo 2 o __ : 64 1 o __ 2 o __ 1 o __ 2 o - __ :: 66 1 o __ 2 o --. 67 1 o . __ 2 o __ ` 68 1 o __ :, . 20 2 0 : ----69 2 0 __ ~: .
t 7 . 1 . o ~ 2 0 __ `~ 71 1 . o -- ~ -2 o __ -~

: 2 0 ----73 1 o __ 3 2 0 __ 74 1 o __ 1 1 o.4 mm 2 0 __ 76 1 o __ 2 0 __ . 77 1 o 2 0 __ 78 1 o --` 40 2 0 --: 79 1 o - __ 2 o __ i 80 2 o __ -81 1 o . __ - 2 o -- :
82 1 o __ 2 o __ Two of the broken transverse rupture SpecLmen9 from the remaining samples o~ each billet, Sample 8, were randomly ` .1 , \ 105Z984 ~

elected for micro-honogeneity. These micro-homogeneity sample~ were polished and photomicrographed at 900X magnification. Tho results t~bulated in Table V below indicate that the averagQ largest titanium-`
`.
. , carbide agglomerate wa~ 12 microns, and thQ average titanium carbido -j gr~in wa~ 4.ô2 micron~. It should be noted that agglomerate~ are ? ~
combinations of two or more grains lnto one ma~s.
', ................................................................ ~. .' TABLE V. BILLET MICRO-HOMOGENErTY j .~ . , '~. ..
~ LARGEST TiC TiC AGGLOMERATES ~
; AGGLOMERATE OVER 10 ~ - ~ ~ -- 10 BILLET SAMPLE EQUIVALENT EQUIVALENT ~ARGESr TiC
. ND. B ~O. ~IAMETER DIAMETE~ GRAIN'~ -;' - ~ 63 . 1 lo ~ 0 ~ ~ 6 ~ ~-~- , 2 15 ~ 2 5.5P ~
`~ `; 64 1 9 ~ o 4 ~ `
2 16 ~ 3 , 5 ~.~ ~-i 65 1 14 ~ 1 5 ~ ~
2 18 ~ 1 5 ~ ~ ' 66 1 9 ~ o 5 2 7 ~-~ - -5-67 1 14 ~ 2 - - 5 `
~ ~ 2 16 ~ 2 4 i` 68 1 14 ~ - p ~ ~
2 20.5~ a ~4 P ~ :; :
69 1 12 ~ 5.5~ . ~
` 2 12 ~ ` 5.5~- ~`
~` 70 1 8 ~ 0 ,, 5.5P .
2 15 ~ 2 ~ ;5.54 ~ ~ I;~
71 1 14 ~ 1~ 4 ~ -`
,~ 2 14 ~ 2 - ~ - t' ~ " ,.
i$ . 72 1 ` 9 ~ o~ 4~ ~ - - -. r3 1 12 ~ 1 3.3P-~ --3 2 15.5~ 1 ; 6 74 1 8 ~ ~ o - -- 5- ~-~ 2 10.5~ 1 ~ 5`` ~`-;~j 75 1 11 ~ 2 ~ 4`- ~ ~;
2 19 ~ 3 -~ -5.5~ ' ; 76 1 175 ~ 5~ `
: -~ 7r 1 lO ~ o 5.
~` 2 10.5~ 1 r ~ l 8 1 9 ~ o ~ 5 ~
~" 40 2 10 ~ O ~ 6 ~ , 79 2 12 2 O 5 ~ ~ ~
1 13 ~ 2. 3 ~ !1 2 17 - ~ 5 81 1 10.5~ ~ 1 - 4 - . 2 12 ~ 2 5.5 82 1 15-5~ 2 ;, 2 15 AVERAGE 12.0~ 1.08 4.82 - 27- -- - !`

.. . .
... . - ~`" ' !- .

lOS'~984 A scanning electron microscope indicate~ that the alumina grain size of this material is of the same order as the grain size (003 - 105 ~) of the sintered alumina alone.
The TiC, however, is on the order of 1 ~, which was the size of the ball milled titanium carbide powder.
As described, the process produces a significantly im-proved product in comparison with the prior art. The increased density of alwmina-titanium carbide indicates that the applied rate controlled sintering technique immediately following the "break away" point maximizes the densification of the material relative to that which was heretofore obtainable. This pro-cess, moreover, i8 appllcable to other powdexed materials once the "break away" point is determined and the inherent or natural densification rate for the material in question is est-blished.

:i . ' ':

; ' ~

. . .

.
' .

Claims (4)

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE PROPERTY
OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. A process for densifying a metal oxide powder into a fine grained sintered ceramic comprising applying a compressing force to the powder, heating the powder at a predetermined temperature rate during said force application to an onset of powder shrinkage, and subsequently densifying the powder through at least two sequentially decreasing densification rates by variation of said force.

2. A process according to claim 1 wherein one of said densification rates is constant.

3. A process according to claim 1 wherein one of said densification rates is non-linear.

4. A process for forming a fine grained sintered ceramic from a metal oxide powder comprising the steps of heating the powder at a temperature rate of 400 to 1000°C per minute, applying a constant physical pressure to the powder while the powder is being heated at said rate until the powder reaches an onset of powder shrinkage, subsequently sintering the powder through at least two sequential sintering rates and controlling the rate of sintering by adjustment of the applied pressure.