| 1 ‘ H {H ‘\ ( ' 1“! W h l '1 ? 7—77-74 P- ____ *7 w '1} W t ‘ W ‘1 MAGNETEC EMEBACTEGNS BETWEEN COSSm ME} ERGN COBALT FLNE FARTECLE COWGSE'EEE THESIS FOB TEE DEGREE OF M. S. IflCHlGAN TATE UNIVERSITY MANFRED 1303138 29:71 LIBRAI‘Ai Y ' Michigan State University Wum_.p. -...o~.—- w... ...4 ' ' '_A""~"'~’.-19<1 \ H1 fl? TX? ALL MODELS COHERENT CURLING FANNING Figure 2 - Schematic Representation of Magnetization Reversal Configurations. The Cross Sections Show Only the Transverse Flux Components. 8 The intrinsic coercive force for the curling mechanism can be expressed as, Hci = 27 K A/b2 MS - Na Mg (2) where A is the exchange constant, b the semiminor axis, and K a shape factor constant varying from 1.08 for an infinitely long cyclinder to 1.38 for a sphere. In practice irregularities in the shape of particles which deviate widely from ellipsoidal or rod shape give rise to more nearly a fanning mechanism and can be expressed by the relationship Hci = Lfar/TMS (3) This is the best expression for materials such as Alnico where the precipitate particles are peanut-shape and have many cross ties. It has been shown that the behavior of specially grown elongated particles having diameters less than 1000Ao agree very well with the curling mechanism ex— pression for Hci~ Crystal anisotropy effects are intrinsic but are not fully understood. The electron orbits in these materials are coupled to the crystal lattice, causing the moments to align themselves along certain axes. For example, barium ferrites and new cobalt rare earth magnetic materials have hexagonal crystals in which the C axis is the preferred magnetic direction. The coercive force of these materials can be expressed as, HCi = 2K/MS (4) where K is now the crystal anisotropy constant. The corre— lation between the calculated value and actual measurements agree only qualitatively and deviations are expected to result from compositional fluctuations, defects in crystal structure causing lowering of K locally, or positive or negative contributions from other anisotropies. As with the shape anisotropy materials, partical sizes equal to the single domain will produce the largest Hci values and be the most resistive to demagnetization. The reversals that take place in these domains can readily be seen by drawing complete hysteresis loops or partial demagnetization curves as shown in Figure 3. 8m Bog/’7 I N DUCT! ON lg; o COERCVWE FORCE A2 __,,,,//”A Figure 3 - Hysteresis Curve for Permanent Magnets -8", r 10 A magnetic material just freshly prepared would have a net magnetic field of 0 until an external field is applied. When the field is applied, any domain walls would move first causing rotation of some of the moments. As more field is applied, rotation of moments in the domains takes place until they are aligned in the direction of the field. The field energy required to complete this alignment is called Bis' the saturation induction. Removal of the external field will cause the remanence to drop to Br, which is called the re— sidual induction of the magnet. Br is directly dependent on whether the preferred shaped or crystal particle directions were lined up with the direction of external field magneti— zation. A measure of the degree of this alignment can be expressed in the ratio of Br/Bis- (5) A value of 1 would represent perfect alignment of the particles with the external field. This value is never obtained in practice because of the difficulty of getting all moments to rotate fully to their preferred directions. This is in part prevented by the inability to obtain perfect shaped particles or crystals. When the external field is reversed, the demagnetization curve Al is obtained in the second quadrant showing how the moments are being reversed. When the remanence (B) equals zero, the Hci value of the material is obtained. Further 11 application of the external field will completely saturate the magnet in the reverse direction, as shown in the third quadrant of the figure. If the external field is reversed again, the hysteresis curve A2 will be drawn and end at the original Bis point. To make the curve completely symetrical the external field must be large enough to fully saturate the magnet at both ends. The curves for crystal anisotropy and shape anisotropy materials are very different in practice for hard magnetic systems. In shape-anisotropy materials, HC and Hci are nearly the same because of the inability to make perfectly shaped particles. In crystal anisotropy materials the Hci is usually much greater than HC, (Figure 4). ’ "' Br LIVER/£515 _ _ _ ,3 , P r / v. Wax I I I I I / 0 3 i ' *BHMGx U I 9 3 k I Q ’ u ' 3 I D Q I I 3 I 4 I l l I“; Ha coe'Rcu/E FcRCE COL-”RC u w: H“ ”c FORCE Crystal Anisotropic Shape Anisotropic Materials ' Materials Figure 4 - Second—Quadrant Demagnetization Curves for Shape and Crystal Anisotropy Permanent Magnet Materials 12 The Hc curve is obtained by taking each Bi value on the intrinsic curve and subtracting the reverse applied H. The crystal anisotropy material "true" curve is usually a 450 line where HCQ: Br, and the energy product BHmax lies half- way between the two points. In shape-anisotropy materials, HC is always much lower than Br, and BHmax occurs at the knee of the curve. To obtain the highest saturation magnetization for any particular material, a high volume fraction of magnetic- shaped precipitate or crystal particle must be held in a nonmagnetic matrix. Thus packing becomes a very important consideration whether the magnet is made by cast—precipitation techniques (as in Alnicos), order-disorder reactions (as in CoPt), sintering reactions (as in ferrites), or by synthetic methods where the magnetic particles are first made and then mixed with a non-magnetic matrix material (as in Lodex). Two extreme behaviors have been observed and are shown in Figure 5. In all cases Bis increases linearly with increasing packing fraction p. The intrinsic coercive force remains constant for crystal anisotropy materials but decreases for shape anisotropy materials. Née1(12) derived an expression for an isotropic assembly of elongated particles whose moments are parallel and l3 ’. o \ E _ Heme.) p=0 ‘ ' a 3. '5 £0 CRYSTAL «0.8 é Q ANISOTROPY o 1 fiP 4:045".l 6 _ '-2 3.x. B"/8'5”” - t" ‘34 .47? S E s 9 ‘~ t 0‘ .2 g w 0 0 PACKING FRACTION 20 4o SHAPE A *- ANISOTROPY .. M Q m— m :- ‘é‘ _ . o; 512 r- =Bis u: 9 § Q .. 20 0‘ A 2 a b q i a I U) Q Q: '" V 3 « I0 -<~+ ~ ”c5 12’ 00 GL2 J ‘ 7000 0.4 006 o PACKING FRAC TION Figure 5 - Theoretical Effect of Packing on an Aligned Sample of Fine Particals with Crystal and Shape Anisotropy (Courtesy of Parker and Luborski 2 ) l4 magnetization is uniform, HCi = Hci, p = 0 (l-p). (6) Later it was shown that this expression would also hold for a random assembly of parallel infinite cylinders reversing coherently. The graphs in Figure 5 are based on the assumption that all particles are perfectly aligned and the domains all are equally magnetized. This perfect alignment and magnetization does not happen under experimental conditions except at lower values of p. The purposes of this study were to determine whether CoSSm would follow the theoretical crystal anisotropic curve and to determine the magnetic properties of mixtures of crystal and shape anisotropy materials. EXPERIMENTAL APPROACH (l3) Julien and Jones reported on the affects ofCX.—§f phases in Alnico and showed how the properties of the magnet decreased as the amount of 1” phase increased. A pronounced “kink" was observed on the demagnetization curve when about 12% 1’ phase was precipitated. This paper stimulated a personal interest in determining if magnetic materials would react in a similar manner. A decision was made to test the effect of mixing a shape- anisotropy magnetic material such as Lodex with a crystal- anisotropy magnetic material such as Co5Sm. Since "kinks" occur in the demagnetization curves of Alnico, a similar reaction might be postulated for the Lodex and Co5Sm system. A second purpose of the experiment was to determine what effect Lodex would have on remanence, intrinsic coercive force, and energy product of COSSm. ‘Lodex and C058m were chosen specifically because their magnetic properties are widely different and such large differences might result in more pronounced particle inter- actions and thus be more easily observable. Lodex, being a 15 l6 magnetic material in a lead matrix, also allowed the two magnetic materials to be compacted without the use of additional binder. Lodex is an iron-cobalt magnetic material electro— plated in a cell having a mercury cathode and an anode of iron and cobalt. The iron cobalt single-domain particles are formed in the mercury cathode. This mercury slurry later is heat—treated and matrix additions are made to obtain specific magnetic properties. It is preformed into an aggregate billet in which the single—domain (50AO diameter x 7000A length) particles are aligned in one direction. The billet is then ground to a size which can be used for final pressing. The Lodex process is outlined in the block diagram of Figure 6. The alloy Co5Sm is produced by melting pure 99.9% cobalt and samarium in a vacuum induction furnace. The billets from this melt are crushed in a jaw crusher and further refined to a 12-micron size in a nitrogen gas-jet mill. Two compositions are made: one of 33% samarium content and the other a 60% samarium-rich phase used as a liquid- phase additive. The two compositions are blended to produce an average composition of 37% samarium;and the blend is placed in a rubber tube and taken to a 60,000 oersted super- conductor where the material is aligned and pressed l7 woavwmm ucmumm m: I mmoooum mcflnsuommscwz xmwoq mo EmHmMHQ onHm.I w musmflm 0253?. ‘42: 024 203< 0252.3 I . ._ o IIIIIIIIIIIII mIAiIIMIIIIIM. hIlIILI .I I III II IIIIII IIIIIII II./.I r..III.IIuIIlH lllll IIIIA II IHIIIIIHI II. III I\ / i I / 20.23.55 I EDDU<> mmwam 0.24 ZO:<. x.mh<2 I H .z 9.11 I .\ . IIIII.‘ . 1i a>0¢0 503.30.300thde— . .. fl .. mag ”$92: fillJ ao— Joood. . ( coo- l8 sufficiently to hold the material together before final pressing. The powder next is hydropressed at pressures of 200,000 psi to a density of 77 - 80% theoretical. After removal from the rubber tube the billets are sintered at 11200C in an argon atmosphere for 30 minutes to achieve a 90% density. The block diagram of the process is shown in Figure 7. The experiment was started after obtaining the desired preformed Lodex and sintered Co Sm. The chemical and mag- 5 netic properties of these alloys are shown in Table l. The Lodex billet was ground into powder using a Weber Boss pulver- izer mill and screened to obtain the +200 —18 Tyler fraction. The sintered COSSm billet was ground in a jaw crusher and further reduced in a pulverizer to achieve a particle distribution as shown in Table 2. By grinding the sintered CoSSm billet to this size powder an aggregate particle containing many crystals is obtained having the preferred magnetic axis aligned in the same di- rection. This in turn assures a maximum moment for field pressing. However, the main reason for making the large particles is to produce a stable powder resistant to oxi- dation; particles of C058m in the lO-micron range are very susceptible to rapid oxidation in air. l9 Co Sm Vacuum .‘ Jaw Pulverize Cast Crushed and Jet Mill 12 M S ize Base Alloy Align Blend "————-60%»R1Ch Material I in f—4 60,000 Field Hydropress ,,,, Sinter ____..I Final Grind Figure 7 - Flow Diagram of C058m Manufacturing Process 20 Table 1 — Lodex and C05Sm Chemical and Magnetic Properties Material Composition Wt. % Lodex 14.1 21.4 5.09 55.1 .91 3.12 .28 C05Sm* 37 63 Magnetic Properties Br Hci BHmax Lodex 6500 900 2.2 C05Sm* 7890 17,600 14.8 * - The magnetic properties of C05Sm are those obtained after sintering the aligned powder to a density of 90% and magnetizing the billet in a 60,000-oersted field. Table 2 — C05Sm Powder Distribution After Grinding the Sintered Billet Screen Size Weight Wt. Percent Microns gms -61 11 2.08 +61 -104 70 12.8 +104 -147 122 22.3 +147 -208 176 32.2 +208 -295 126 23.1 +295 -417 33 6.05 +417 8 1.47 21 "As sintered" CosSm properties are shown in Table 1 after being magnetized in a 60,000—oersted field. After obtaining the desired magnetic powders, mixing experiments were carried out for various combinations of C058m and Lodex. The powders were accurately measured to the nearest 0.05 gm on a Mettler P1200 balance. It was found necessary to add 0.8 wt% of MoS2 as a lubricant to prevent laminations in the samples after pressing. Other lubricants probably could have been used, but this is a convenient one for Lodex die—pressing. The weighed powders were placed in 2-oz glass jars and rotated end for end at a speed of 12 rpm for 15 minutes to achieve blending. A stainless steel nonmagnetic die was made that had a cavity % x 3/4 x 1% inches with outside dimensions of 3 x 5 x 1% inches. A picture of the unassembled die is shown in Figure 8. Then 6 grams of mixed powders was poured into the stain- less steel die and placed in the magnetizing coil (Figure 9). This coil was equipped with tapered pole pieces that con- centrate the magnetic field to 11,000 oersteds in a two inch diameter cylindrical region between the pole faces. The die was placed between the pole pieces so that the die cavity was in the center of the pole—face region. The magnetizer was switched on; and the powder was aligned in the 11,000 oersted 22 Figure 8 — Nonmagnetic Stainless Steel Die and Punches Figure 9 - Electromagnet System for Alignment of C05Sm Powder 23 field. While the field was on, a pressure of 1300 psi was applied by the vertical ram to prevent the aligned particles from becoming random again when the field was removed. The die then was removed from the magnetizer and pressed in a Pasadena hydraulic press (Figure 10) to a pressure of 210,000 psi. Figure 10 - Hydraulic Press for Final Pressing Cycle After removal from the die, the sample dimensions were measured and densities determined. In order to achieve reproducible magnetic measurements, the samples were all {73" 'h 24 saturated at 60,000 oersteds in a General Electric Model 65E-175 amp superconducting magnet (Figure 11). The pressed- magnet size was small enough to allow the magnet to be inserted in the field and pulled out again without having to turn the field off. Fields of 25,000 oersteds which could be obtained from an electromagnet were not sufficient to saturate the composite materials. Figure 11 - Superconducting Magnet Final magnetic measurements were made on an O. S. Walker Model MH 5.0 Hysteresisgraph capable of measuring coercive forces as high as 25,000 oersteds (Figure 12) and of drawing 25 curves on a X - Y console plotter. Accuracy of the electronic integrating flux meter is 1% or better, and the Hall—effect gaussmeter is %% or better. Figure 12 - 0. S. Walker Hysteresisgraph Curve Tracer Figure 13 shows a block diagram of the test features of this equipment. The Hall probe measures the coercive force of the applied field. The Bi coil is placed around the sample. This coil is composed of two components; one measures the B or induction of the sample, and the other part measures the H of the applied field. Since(l) Bi = B + H (7) the resultant from these coils measures Bi directly. The Bi coil consists of 490 turns of 3—mi1 wire having an overall resistance of 53.4 ohms. The effective area of 26 mun—800mm > IX > X TV Emummfla xoon HmcoHuUCSm Eoumxm I ma wusmfim L at... 5:33. «0:902 III >4..an 1... "too ¢u>>0a o n 5 2652 T Saga“ 1 :ou a a :I ,3 L :25. dongOu . (H Juumu unwauunmim 3.6 Faintinlx J I 523$ uFiLVuom— 315 y La g; g 3:6..in I .0 \ _ UIVU \ F I :1 fII\\ 555:3: 2205.”:3 a . «mums. omhwfio .__._<: elIIIIiu “.5528: - . . 322.3... coor— o.o..\ \ 27 this coil is 913.4 cm2. For measurement of Hci of these samples, a 290.3—ohm Hall probe capable of measuring up to 10,000 oersteds was used. Output from these probes are fed into their respective amplifiers so the Bi and Hci can be displayed on meters or directly plotted on the X4Y recorder. This equipment has a balancing circuit which allows stabilization of the electronic circuitry to prevent measurement drift. To measure Br the sample was placed between the pole pieces and the Bi coil placed over the magnet as shown in Figure 14. Figure 14 - Closeup View of Hall Probe, Bi Coil and Pole ' Pieces 28 The gap between the pole faces was closed until the magnet touched both poles. The recorder was balanced to read full scale when the magnet was in this closed magnetic circuit. Next the magnetic circuit was broken by opening the gap between the pole pieces. The flux measured at this point is the leakage flux from the magnet. Removal of the coil from the magnet causes a further change in flux measurement (Y direction) which represents the remainder of the induction or remanence of the magnet. The summation of these two values equals Br (see Appendix A). The Bi coil was placed over the magnet again, and the pole pieces were closed. Adjustments were made on the X - Y recorder, and the pen was set at the established Br value. By turning the H—drive control, a magnetic field was established counter to the field direction of the magnet. The Bi and Hci values were recorded on the X — Y recorder. Only the second quadrant of the hysteresis loop was drawn. This equipment could not be used to draw a complete hysteresis curve since insufficient magnetic field was available to saturate these magnets in the reverse direction. Measurement errors were found to be less than 2% provided the samples were remagnetized each time at 60,000 oersteds in the super— conductor. EXPERIMENTAL RESULTS Table 3 is a summary of the magnetic data obtained for various compositions of Lodex in C05Sm. Figure 15 is a graphical representation of how Hci varied with various amounts of Lodex. Excellent data points were obtained on the Lodex—rich compositional side. The scatter obtained on the CoSSm—rich region was due to the poor structural strength of the compact. Corners of the magnets would break off easily with compositions containing less than 14% non-magnetic binder material. In order to reduce the scatter and obtain some repeatability, the samples were sprayed with lacquer after die—pressing. Remanence (Br) for Lodex—C05Sm compositions is plotted in Figure 16. The data shows remanence increasing as the Lodex content increases. From theoretical calculations (Appendix A) the Br for the 100% C05Sm compact should be 6950 assuming perfect alignment and stress-free surfaces. Only 85% of this value was obtained in the experiment. It is relatively certain that some misalignment does exist since cogging of particles must take place during field 29 .mnmumumo 000.00 um UmNHumcmmE Ucm Hmm M OHN 02m mvmumuwo ooo.HH um Ummmmum cam cwcmflam memEmm 3O 0.¢0 0 0 0.00H 0N.N 000 000 0000 NH 0.00 0.0 N.0 0.00 0.N 000 050 0000 VN 0.00 0.0 N.0H 0.00 0.N 00HH 000A 0000 mmm 0x000 mm 0.0% 0.0 N.0N 0.05 v.m 00¢H 00NH 0000 mm 0.0m 0.0 N.0m 0.00 0.0 000d oova 0000 mam 0.00 0.0 N.0m 0.00 0.N 000a 00¢H 0000 HN 0.0m 0.0 N.0v 0.00 00¢N 0000 MON 0.mm 0.0 N.0v 0.00 0.0 000m 005a 0000 ON 0.0m 0.0 N.00 0.0v 0.0 000m 000d 0000 0H 0.00 0.0 0.00 0.00 0.0 000% 00¢N 0000 0H 5.0a 0.0 0.05 0.0m ¢.¢ 0000 000m 0000 05H 5.MH 0.0 N.05 0.0m 0.0 0050 000m 0000 5H 0N.5 0.0 N.00 0.0a 0005 0000 00H 0N.5 0.0 N.00 0.0a 0.0 0000 0000 0050 0H 0.0a 0.0 0.00 N.¢H 0000 0000 Nlm0H 0.0a 0.0 0.00 N.¢H 0000 0000 Himma 0.0a 0.0 0.00 N.¢H 00.¢ 00H0 00am 0000 00H 0.0a 0.0 0.00 N.¢H 0.v 0050 000m 0000 0a 0.0 0.0 0.00 0.0 N.5 0005 000v 0000 00H mxoum 0H 30000002 .900 Emmoo x0003 xmemm 000 00 H0 .02 ofluwcmmz X .03 R .03 X .03 meEmm 002 .x. .03 mqu mnmEEsm xmwoq I Emmou I m wanna 31 muummEoo xwvog I Emmoo CA wouom mSAUHwOO oflmcfluucH ammoo 2H xmooq 0200000 .03 Om 0.0 05 On 1 4 Oh 0? On A q mo coflumflum> I ma wnsmflm m V. spaiglao {or X TDH V0 32 II.0.U\ . Do. muommeou ommmmnm xmvog I Emmoo CH coHuUDUCH Emmoo zH xmooa 9200000 .03 0Q 05 00 0.“ Q? 00 < 4 1 a q 00 cofluMHum> I ma musmflm ssnvs EOI X 18 33 pressing which would prevent some particles from orienting completely in the proper direction. A further indication of misalignment is the scatter of points in the graph. What effect surface damage has on the ground CoSSm aggregates was not investigated in this experiment. As predicted, ”kinks" did appear in the demagnetization curves for certain mixtures of C05Sm and Lodex. Figure 17 shows typical intrinsic curves for 0, 30, 60 and 100% Lodex mixtures. These "kinks" were observed in mixtures containing from 20 to 80% Lodex. Figure 18 shows how energy product varied with Lodex composition. The dashed line represents the region where the maximum "kinking” occurred in the curves. Since BHmax is the maximum area under the B—H curve, any "kink" in this curve would lower the BH value. To assure that the "kinks" in the curves resulted entirely from the interaction of the FeCo particles in Lodex and not a dilution affect, mixtures of C058m and pure lead were measured. No "kinks" were observed in the demag- netization curves. Figures 19 and 20 demonstrate how Br, Hci and BHmaX varied with lead composition. 34 Emmoolxwvoa mwmmzm ofluwcmmz 039 mnu mo uowmma mcu 0cfl3ocm mm>usu coflumufluwcmmEmQ h 5H wusmflm xmooq nx03 gm 00 2H xmooa R00 0 w w m 00:. .51» o N v w 0 002.3% + q . A _ _ _ 0 ~ 1 q .1. 1 L .U m- r 50000 7: 000.003 Rom Emmoo a 00003 R0 W m AmQ\X .541» G N V W I m KMQ\ X “Qty ! d s O u a di - q ‘ a .1. ‘ 35 muommEOU memwum 80000 CA Axmfimmv posnoum mmuwcm mo COADMHHE> I ma musmflm Emmoo zH xmooa 9200000 .03 0.0x 0% 0% Oh 00 0.0. Q? 00 ON 0‘ 0 1 q l 4 q (00.60111) XEUIHH 36 000 0cm 00 mo c000oc5m 0 mm 9m 00 m0sDOE< msoflum> msam Emmov I ma m0500m Emmoo CH x0000 0cwouwm .03 06\ 90 00 68 00 06. 0% an gm 0‘ o o / a q q i . q 1 . q 0 / / / \I / /. /. NI / I / m I H o m- r. $. X Tl. X OE v: I.“ O a 1 3 1| m‘ W m s 0- N. o 5' .0 In 37 m0ommeou memwum QmIEmmoo :0 Axmemmv 0oscoum >000:m mo c0000000> I Emmou :0 @000 0cmouwm .03 so, am oh 0m 9w 2.. 6m 6.... S — q a q a q q q 4 om musmfim (ooobom) XQmHH DISCUSSION AND CONCLUSIONS A previous study of two-phase (o( —‘r) Alnico 8 alloy materials reveals the presence of a bend or "kink" in the demagnetization curve, after a pre—field treatment isothermal 9000C temperature. The reasons for the occurrence of the "kinks" were unexplained. From this study, it was postulated that similar "kinks" might be found in the demagnetization curves of composites made from crystallographically—aniso- tropic and geometrically—anisotropic magnetic materials such as Cossm and Lodex. In the experimental results shown here, indeed such "kinks" are observed in the demagnetization curve of a com— bination of these two magnetic materials in the range of 20 — 80 weight percent Lodex. This is found without the need for heat treatment, as was the case with the Alnico 8 alloys; however, in the C05Sm—Lodex composites the magnetic phases are already present and being held together with a lead matrix. A possible explanation of these "kinks” may involve an early flipover of domains of the lower Hci Lodex material until all the domains are turned, while the remaining 38 39 demagnetization curve characteristics are solely those of the higher coercive C053m material. This would not explain why Hci is still changing at either end of the "kinking" range. The intrinsic curve on the Co5Sm rich-composite varies linearly with Lodex additions; whereas, the curve for Lodex rich composite does not vary linearly. From theory(2)if a crystaleanisotrOpy material such as Co5Sm is diluted with a magnetically inert material, the Hci should remain constant; the shape—anisotropy material such as Lodex, on the other hand should vary linearly under similar dilution. Some interaction mechanism must occur between the magnetic particle moments in addition to dilution material effects throughout the compo- sitional range. Another indication of the presence of some interaction may be found in the slight increase of remanence (Br) in the C05Sm—Lodex mixtures as the Lodex content is increased, in spite of the fact that the nonmagnetic content of the system changed from 22 to 64.5 wt. percent. In this set of experi— ments, the 100% Lodex material did have a higher Br than the 100% C058m magnets, yet one would expect the Br of either material to decrease with dilution by the nonmagnetic (2) material. 40 To further substantiate this premise, various compo- sitions of C05Sm—lead magnets were made. The value of Br for these magnets decreased with increasing amounts of lead; whereas, the value of Hci remains nearly constant. These results, Figure 19, are in reasonable agreement with Neel's(2) prediction for crystal—anisotropy materials. The observed slight deviations from the theoretical curves may result from imperfect alignment of particles or from crystal imperfections. Based on a comparison of the results for COSSm-lead with those of C05Sm—Lodex, it appears that some interaction between the magnetic particles must be occurring. This effect should be examined further to determine the exact mechanism, but such a study is beyond the scope of the present thesis. Although the Br remains almost constant and Hci increases as the Co Sm is enriched in the composite systems, no striking 5 increase in the maximum energy product occurs until 80 wt.% Co5Sm is reached. Since BHmax is obtained by plotting the product of the coordinates of the demagnetization curve, the values are directly affected by the ”kink” in the demagneti- zation curve. Thus very little increase in energy product will result from increase in HC as long as "kinks” occur in this curve . The physical properties of the pressed part exhibited 1‘11- 4 1 ROBERT Sum" poor mechanical integrity until the lead binder content reached 14 wt. percent. Normal handling of the specimens resulted in edge breakage; consequently, there was considerable scatter in data from the Co5Sm—rich composites. In conclusion, the data clearly shows that C058m and Lodex cannot be mixed without generating "kinks” in the demagnetization curve. Whether composites of two or more shaped-anisotropic or two or more crystal-anisotropic magnetic materials held together with a nonmagnetic material would exhibit similar behavior is not known and would be a fruitful subject for further study. BI BLIOGRAPHY 10. ll. 12. l3. l4. BIBLIOGRAPHY R. J. Parker and R. Studders, "Permanent Magnets and Their Applications", John Wiley & Sons, N. Y. 1962. F. E. Luborski and R. J. Parker, GE Report 66-C-252, October 1966. F. Westendorp and K. Buschow, Solid State Commun., 7,639 (1969). D. Das, IEEE Trans. Magnetics, MAG—5, 214 (1969). D. L. Martin and M. G. Benz, GE Tech. Report 70-C—261, August 1970. D. L. Martin and M. G. Benz, Ternary Cobalt - Rare Earth P.M. Alloys, 1970 International Conference on Magnetism, Grenoble, France. W. D. Kingery, "Introduction to Ceramics", John Wiley & Son, N.Y. 1960. J. Frenkel and J. Dorfman, Nature 126, 274, (1930). D. Hadfield, "Permanent Magnets and Magnetism", John Wiley & Sons, N.Y., 1962. F. E. Luborski,_J, Appl. Phys. 32, 1715 (1961). F. E. Luborski and E. R. Morelock, J;_Appl. Phys. 35, 2055 (1964). L. Neel, Ann. Universitinrenoble 22, 299 (1946). C. A. Julien and F. G. Jones, Cobalt 21J(June 1965). Manufacturing Methods and Technology for Processing Cobalt-Samarium Magnets, Report TR-612-9A, WPADC, Contract # F33615-70-C-1098. 42 APPENDIX APPENDIX A Theory and Calculations For Magnetic Measurements Br Measurement In measuring a magnet open circuit, one cannot measure the Br directly unless the length of bar is infinite. The magnet will operate somewhere down the curve depending upon the length-to-area ratio of the sample. [3" ‘—"“"3h1 l‘b I ’0 I I, 1. i 9 I ”c Hn1 Figure 21 — Curve Depicting Open-Circuit Magnet Conditions Applying Parker and Studders analogy(l) for unit prop- erties and geometry relationships for magnetic materials, gives, LmHm - LgHg = 0, (Al) where Lm = magnet length, Hm = magnetization potential of magnet per unit length, L9 = air-gap length, Hg = air-gap 43 44 unit potential. If the air gap and magnet neutral section are equated, and if it is assumed for the moment that all magnetic lines reach the air gap: where Am = magnet area, B magnet unit density, Ag = m gap area, B9 = gap flux density. By combining equations 1 and 2, Lm = LgHg/Hm, Am = Ang/Bm, _ 2 vm — LgHg Ag/BmHm (A3) where Vm = magnet volume and Bg = Hg From Ohms law and using magnetic circuit analogy Flux = Magnetomotive Force divided by the Reluctance, it can be (1) shown that the "load line” as shown in Figure 21 is equal to, Bm/Hm = LmAg/Ang = Lm(l)/Am(R) (A4) where R = reluctance. Thus for a specific length and area of magnet one would measure the point Hm and Bm when placing a B coil over the magnet while having the magnet open circuited. By placing the magnet between two pole pieces with a magnet return path, the magnet load line will recoil up the demag curve to the Br value provided the magnet L/D ratio does not put the magnet below the knee of the curve. If the load line is below the knee of the curve, the magnet would recoil at a lower value '45 than Br° One can calculate Br from the open—circuit conditions once you know the Bm and Hm points. This is often hard to do since the exact load line is not known until the reluctance value is obtained. By using the integrating flux meters in the O. S. Walker equipment, a relationship is developed de/dt = KfE dt = B (A5) where t = time, B = flux density, E = voltage measured by integrator amplifiers, K = constant for the particular sample (C.F.). This relationship is shown in two parts in Figure 22. ,_1 a (fruit .... .éu I T 2 I 3 I :3. Kfi-‘zdt E ' I _ [Hm _ 7 fCC‘ERCI't/E [rcgcg' Figure 22 — Curve Showing Integrating Flux Regions By opening the pole pieces and getting the open—circuit conditions for the magnet, the O. S. Walker equipment will measure the first anldt. When the Bi coil is removed, the 46 second integralJEzdt is measured. By adding the quantities KEfEldt+ [E203] =Br the remanence of the sample is obtained. Coil factor (K or C.F.) Calculation for the O.S. Walker 4 . . . - C.F. = 10 x number of turns in Bi c011 x spec1men area x 10 3 102 C. F. for this experiment = 49 x specimen area in (cm) Theoretical Value for Br for C05Sm Powder From theoretical relationships (Figure 5) Br is directly related to the density of the compact for shape anisotropy Br sintered = density sintered Br at compact density density compact Br sintered for Co5Sm material used = 7890 and density = 7.37 gm/cc 7890 = 7.37 where Brn = Induction for powder compact Brn 6.5 Brn = (7890) (6.5) = 6950 if aligned perfectly 7.37 The measured Br for the 100% C058m compact which was die pressed at 210 K psig = 5900 gauss 5900 _ .85 = Fraction of sintered Br obtained 47 48 0H.m @000 .> 00.mH mmm. 005. 00m. 00m. com 000 cc00 NH mm.m @000 .> ca.5H m0m. 005. 00m. mam. c0m c5m coco 0N 05.0 @000 .> c0.5H 000. m05. 00m. 0am. ccHH coca cccw mmm wxonm mm 50.0 0000 5m.mH 050. 505. 5mm. 00m. cc00 coma c0c0 mm 00.0 @000 0m.0a 500. 005. 00m. 0mm. ccmH cc00 c000 mam 00.0 @000 0m.ma 500. 005. 00m. 0mm. cmma cc00 coco Hm cc0m coco mom 05.5 0000 05.ma 000. 005. 00m. 00m. comm cc5H coco cm 50.5 0000 00.cm 0H0. 505. 00m. 00m. 0000 coma c000 ma HH.5 0000 00.0m mm0. 505. 50m. 00m. 0000 cc0m c000 00 00.0 0000 0m.mm 000. 005. 00m. 05m. ccm0 ccmm 0000 m5H 00.0 0000 cm.am 500. 005. 00m. H5m. 0050 c000 coco 5H 50.0 000m mm.mm 500. 005. 00m. 00m. ccc5 comm Mod 00.0 000m mm.mm 000. m05. 50m. me. cc00 00mm 0050 ma 00.0 000m mm.mm 500. 505. 00m. 05m. c000 ccm0 NIm0H 50.0 000m mm.~m 500. 005. 00m. 55m. comm c000 HI00H 50.0 000m 0m.mm 050. m05. com. 00m. 0.0 cc00 cmam cc00 00H N5.0 000m 00.mm 000. m05. 50m. cmm. 0.0 0050 0000 cc00 0H c0.0 000m 00.mm m50. 005. 00m. cmm. m.5 0005 c0c0 ccm0 m0a mxoum 0H 00\m 000000 NEU .nUcH .nocH .nocH .Xmemm 000 um um # mamamm >00mcmm 300:0000 H000 0004 £00cmq mmwchHzB £0003 mm00momEOU x0000 I 80000 00 m0mo Hm00m>£m cam 0000:002 I 0 mHQmB 49 0.0 0000 0000 0000 000 5.0 0000 0000 0000 00 0.0 0000 0000 0000. 000 0.0 0000 0000 0000 00 5.0 0000 0000 0000 000 00000m 00 05.0 0000 0050 0000 50 0.0 0005 0000 0000 .00 5.0 0005 0000 0000 00 x0500 000 00 0m # 000200 50.0 05.00 000. 005. 000. 000. 00.0 00. 0.0 00 000 00.0 00.00 000. 005. 000. 000. 00.0 00. 0.0 00 00 00.0 00.50 000. 005. 000. 000. 00.0 00. 0.0 00 000 05.0 00.50 000. 005. 000. 000. 00.0 00. 0.0 00 00 00.0 00.00 050. 0054 000. 000. 00.0 00. 0.0 00 000 00.0 00.00 550. 505. 000. 000. 00.0 00. 0.0 00 00 00.5 00.00 000. 505. 000. 000. 00.0 00. 0.0 00 50 00.5 00.00 000. 005. 000. 000. 05.0 00.. 0.0 00 00 00.5 00.00 500. 005. 000. 000. 00.0 00. 00. 00 00 00\0 000000 060 .2000 .2000 .2020 80000 00 02 0000000 0000 0000 200000 20003 000000028 Em Em 90 E0 00 $03 # 000E0m 0000000800 0000 a 60000 00 0000 00000020 000 00000002 I 0 00009 ‘ .— "WWI" r- w rvn‘v . W" fivj va‘ v WWWW""'WJV F7 rv ,w‘r—v“ “T v' ‘w. v W. v— 50 -_—.“ld~.--I'-- »- r ~ .-~.\.--'~IQ.~...L‘ m . ’~.‘...J. 4.-.. . .w .~a..‘c-a‘- .1: “A ‘ a“’. l»' -—c.~.a' ..c-__‘--n‘- )- —-;---~ ‘\.4 :3 \ ”t: "1'27 L/LI'LIJ v U ”55%“ WElGHT Sm En~ou slallfi ...)....--cIOUIII 11...er NOU in! c. )ucu III... N \. III! to 8 1 3500 I400 PN\ .. I n l- o‘v‘v. a I n...- ‘llv|‘n.')\‘l‘!lvl a $ a. m 5 z 7 m m m U . - .H. s I ma. ) . .0 m 5 9 a +6 0 8 o 7 . w m 9 m, m 5 cw. S o a 0 0+ 50/ C C M M O T o A hm Enqou + Samoa 3 ~ ill- c O ulllil¢4.oom..auhcjda.'!. IIIIIIIIIIIIIIIIIIIIIII 2 . Em .3+ :5w 5 II...” H.141. I d. IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII m “Enzoo + A8. m . a O O O O O o w m. w o o o o w w 2 l M, 9 8 1 6 5 4 3 u. .szEmmn—Emh (1968.), ?. 323 1:}. I J’LCM "vfi-‘u- o-v‘“A-w—Ir~, u- nn'vm-mmo - - H" .f-C' - ~§'Ji‘"~’--w*v Figure 23 e Co-Sm Phase Diaggam ’--~*Ql-y~..— . . ‘ -- panm~wgu . 1 . ‘J‘J—z L s_ IA.- . n A Lat-‘1‘. .. .A. a... A.- J‘_._ I'LA-‘L ... AAA Ah- mA: n—ni t—nA“ A-.‘. .- ._ __ _-._ ‘_ —-- . .. N . a...» 51“....1. 341 3i . \ 6‘31“... angrilc .ulrtl . . .Fflolfl... IA.N4EV.i . c. ' “dag—‘O-E‘v 7%...“ M _ _A “‘0.‘ MIC HIGAN STATE UNIVERSITY LIBRARIES uynmu» MI 30713 60 IIIIHILIHH 1293