Document 06EE67k3gBXenGjj3RE6gondx

% *-.? />>i$r rstbs^.^gii^c^OTiicflr'oi'________________________ S%5- Ssjt7- ,*#^ W'V?1 ..-T' -` r, ' Jv 1 - /' -" -*-v `>3 1.' % MagfietK^ Scpmti<t^j^^^;x-<>*^**j|ShS$p?r?>4, A."sjiirr y-'^Lr; 'tt/'i-*.J"S*rt^jSV-'*;; KJt n-fi ** '` ^ /;-.sySsr'*"' **22SJ5 St 5. BADOLLET* ani ..Jit'. Ht . - , t **?*; <;- - -' . .. , ... , , . v . 1 - ':. -' ,^s - >:-.>/ ` , *";; ' * . .->--'- . '. v,. .. `.' ,"; s*'.'r**-'i<. T. i': ". ; " -r .- (Apnnal Gettetai Sleeting, Quebec City, March-, 1961) *. -x^^ " .-'* -''V ' ~y . "y *?;i| agnetite* occurs as an Mimpurity in most deposits of chiysotile asbestos, particularly to mate a* direct separation of mag results.-): It is convenient to describenetite* It .if eaay toseparate a con- . the combined effect of these factors siderable portion of magnetite to by the term "relative permeability." the important ones in Canada, this fishion.but ithaa been found The relative permeability is in- ' This magnetic iron oxide is present too difficult - to remove the last- creitoed if the actual permeability is t in addition to the iron which is sub traces to provide an accurate test. ' large-or if the particles are large, . stituted for magnesium in the crys-... With: the advent'- of' more effective elongated or aligned -parallel to the . tal structure. It is a great disadvan permanent'magnets, efforts of this naguetie field (I, 3^ 4). i; tage in fiber to be used for electrical type hecAmer easier, than with the applications, primarily; because it it electrically conducting and tends to provide a path for breakdown old electromagnets, striiwandecided to mate hiother trywith them. ' There is another reaspn for attempt ltecogniaing. the. importance of , the- difference between magnetic rating and magnetite content, Johns- through the insulating barrier which ing such an extraetiott iu connection ' Mtiitville long ago developed an ex- ~ the asbestos is supposed to provider with an extension of the magnetic tension of the procedure to provide - It is the largest of such particles which are the most likely to provide such a path, being able to bridge more of the thickness of the barrier: To evaluate the quality of asbestos fiber for electrical applications, it ' would be desirable, therefore, to know the number of particles of the various sitics present. However, the magnetite particles adhere tenacious-* ly to the fibers; and it has not been found practical to make a. particle: size analysis as a routine test: of fiber, although it has. been dotm-oCv casionally. Instead, it is usual to de pend on some measure of the guUUp tity of iron present. . - ratihgtest,,1 *- - .y. / IntKe magnetic, 'rating test, 10 . gms of fihez are placed to an to- ductance coilwhicb forms one arm of- : inductance bridge-: Eitbei^ ttitf 'chmu^ to a standard hife dneton rebalance th:' brMgeoi to^4^5Sr,totofrunBals.ucc of iih^ei.:>y- a/gab^' vanoweter itouttoused as is. meas ure a# the"telgnetic;rating;ifefdre Use,tiie standard1 inductor or the galvanometer lit calibrated With a standard sample oif magnetite to make the to^yulnenf'direct reading. de- . an estimate of the relative .perme ability. A comparison sample bf the 1 liiapietito in toe specimen wls ex- * tracted with an electromagnet. The - extraet waa repeatedly ground with a mortor and peagle to free the iCagOetiter from the fiber. This, of course, broke up soma of the par- tldcsi) itt pnrt vitiating the effort ~r '. , by reduemg the relative pertneabil- It^.'.TSe procedure was also very , xc^j; time consuming. Therefore, another incentive to improved magnetite ex- , traction was to provide a better , yv. way of obtaining the comparison. [' sample needed to convert magnetic ?f.' rattog-to magnetite content. * -\,ji , r 3.5h*Iv '' ' ' " ' 3 A chemical analysis; of' 'Ane apiJaratttS was designed: to tity of all forma of iro and therefore ' ptovide a strung magnetic field areK- ^!t'i used, but it roust fee.; "" 'tottog"-dependson scv* '.r a large arCa so fiber etwld be eg' y?' the amount of _______ __ ::$is*rj .... i sddtohto to tbe qnan- potwd-to it effieiently. It consisto> M^ crystal. This .;itpennjeabiJily, of a plastic vessel 6 iii. in diameter.. ;V,^ fiber to fiber .... 'tm ' .............. . e siitd' orientation aud! J2 im high; around. whid can i; ' ror in the of''th^;:.'|(asilcles,' 'The, ...efforts r of , hf phieed a braag frame supporting, rating test is 1ilH^Mp^ use f l)'. theitoviktte* factors ato Snrprisingly IT Alnieo magirets. The magneto It depends on-flpLlSigMSfe effect large: Fot:;e*ample^ the magnetic are to-tbe form of channels, 6' ijn- of the magnetitoii^^fe-if; ignore*, the' rattag .of ite ^bestos paper, will lcmg, with Bides - of opposite mag iron in the crystal. H is morenearly related to the quantity desired-.UnV, fortunately, there is not a one-toone correspondence between ' the magnetic rating and the quantity of : magnetite, and a more straightfor ward test would be desirable. Be cause magnetic forces are fairly eas ily applied, many people have tried ehaiigehyaftoto iOpercent depend- netic. polarity* They form a snug? .1 ingpn wbctfesy tbe papev to rolled " fitting, ring, atound the vessel,' sur- up to Nto s^apto Itolder vrito- the'- rounding the lowef fl in, of if, and , magnetic fteM pSkjtoltd tb the plane : present to it a succeaaioft of alter? r of the sheet w i* -;mit into ssaall : nately north and' south poles, apfe* piece* aitd`|daced iuto t&e. h<daer' stf'1 pruximateiy equally ' spaced around the ficldMi perpendicular to the its- periphery: Magnetic particles to* sheet.*' (Thd p^per-hiaking process an aqueous storry tend to adhere tends to piece :e{aiigated particles, to the watik opposite the spaces bethe usual, fops of magqetitey to the" twequ t^^ polea. When the sep- . * Retired Section Chief, Johns Manville Research Center, Manville, Now Jersey. `Physicist, Johns Manville Re search Center. plane of tlto, jAeet. -Therefore, with, the cut pihees;: alinosk no particlea are parallel'to the magnetic fjeld and a low value for magnetic rating aration ia deemed adequate, theslurry can be drained through a hole in the bottom of the vessel. Then the vessel is lifted ont of the ring of . UCC 0.04632 -- 56-- magnets, and the magnetic material is washed from the walls into a beaker. In use, 20 gms of fiber are dis persed in two liters of water and are brought into the magnetic field at the wall by stirring with a wood en rod. Both the magnetite and fiber fractions are recycled through the separator according to the schedule in Figure 2. The final dry product is weighed and the value ob tained divided by the original 20 gms is taken to be the magnetite content in per cent. Tt soon became apparent that some further agitation was needed to separate the magnetite from the nonmagnetic material. The size of fiber clumps and rock particles which were adhered to insignificant particles of magnetite attested to the strength of th#l$gnettc field provided. Treatm^i^;'for a Waring Blendor at full spggd for 1 minute was found to increase, the yield greatly and seemed to leave the magnetite particles relatively intact, as there was no change in the mag netic rating reading of magnetite concentrate before and after the treatment. This was made a regular part of the process. A series of fibers of Grade 5 through 7 were separated by this technique. The results are plotted against magnetic rating in Figure 3, Magnetic rating was used as a readily available criterion, not be cause it was considered to be a spe cially good one. The early results Mognatlc Rating Figure 3.--Magnetic extract vtrtut magnetic rating -- 57-- j i I nearly equal tp the ehagpetic rating,! This was t&e ease with all' the & grade fibers,. andallbutSof the 6 grade. HowetfsJ^iw*<jral 7 grade fibers, gave. erKWf^i^ttg' largiJ mag netic extadtet;.ji|iB8|BB|.;doable the value ezpecte^y^HKpte| magnetic rating of the this time the study was completisd,ii was ap parent that the high values did not come from a distinct class of freak fibers, hut that there was a continu ous gradation from those whose magnetic extracts were essentially equal to their magnetic ratings. The errors in the separation are of two kinds: some magnetite from the fiber sample is lost and does not end up in the magnetite fraction; and there is some material other than magne tite, principally fiber and serpentine rock, in the magnetite fraction. Be cause of the sources of variation in magnetic rating which are included - in the concept of relative permeabil ity, at least some of the dispersion of the results is caused by differ ences of magnetic rating and not by inadeqnacies of the separation. In formation about the error in both tests can be gained by a chemical analysis of the magnetic extract and by the extension to the magnetic rating test previously mentioned. A chemical analysis is more effective here than for the original fiber be cause in the magnetic extract the fibrous portion has been greatly re duced and the error in the correc tion for iron in the crystal structure is not serious. The actual magnetite in the specimen is estimated as fol lows. . ... . . ... setid rating test, a ' ThlS equation is used' to find the fiber- specimen with magnetite con relative permeability of the fiber tent, MC, and a relative permeabil from the MR reading of the magne ity, fir, will have a magnetic rating, tite fraction, using for Q the weight MR, given by the following rela of the magnetite fraction corrected tion : for the nonmagnetite material in it. The magnetite content is then calcu MR ~v,MC lated from the relative permeability - U) and the magnetic rating of the orig inal fiber by Equation 1. This relation is essentially a defi nition of relative permeability. The magnetite content of the 10 gm specimen, in terms of the weight of magnetite, Q, is: This analysis was made on a new series of 13 fibers' covering most of the range of the early study but with more of the ones which gave a large magnetite extract. The results are shown in Table I. The fact that the relative per MC 100 X Q/10 = 10 Q meability is greater than one for most fibers means that the magne In times of Q, the magnetic rating is tite in those samples has a greater magnetic effect than the standard MR = m,Q used to calibrate the apparatus, probably because the particles are larger and more elongated. Relative In terms of Q, the magnetic rat ing is: MR 10y.r Q However, the value indicated by the magnetic rating apparatus does not depend on there being 10 gm of material present, but only on the quantity of magnetite and its rela tive permeability. Because the term "magnetic rating" carries the im plication of a measure of the per cent magnetite contained in a 10 gm sample, the .term-"MR reading" is used to designate the reading of the apparatus when' any other sample is introduced. Its value is unaffected by inert materials which may be present and. is: ' " permeability is also a measure of the effectiveness of magnetic rating as an estimate of magnetite content, it being the ratio of the two. It can be seen that it is not any too precise. The chemical analysis show*' that the extracted material is only about 70 per cent magnetite for "normal" fibers, considerably less, down to about 45 per cent for fibers which yield an abnormally high amount of extract. The process is somewhat better with regard to loss of magne tite, 80-95 per cent of the total in, the fiber ends up US'the magnetic extract. " " , A question which naturally arises is: How much better would the sep aration be if more effort were ex MR: > lOwQ . (2) pended in loosening the magnetite Fiber GRADE 4T 5K 5M 5R 6D 6D 7D 7M 7M 7RF 7R 7RF 7T Table I Calgulatki#*?o Magnetite Content MC and Accuracy of Estimates ' . - I ...... ' ' ./YfcV" 1 - '. >: Smi. ' Magnetic Extract.r ' CtBEM. . : - - :: Actual ANAL. % MAG MR RDG MAG, Q Rel PERM. t>* Calc. MC MC7 /MC; v-'-t 2.70 2.45 2.80 4.15 >. 4.05 3.8 3.55 70 70 70 78 2.05 1.89. 1.09 2.11 1.29 2.15 1,96 - 1.10 2.23 1.26 3.3 2.84 1.16 3.58 1.13 3.1 2.72 1.14 3.33 1.07 3.65 4.15; 70 3.3 2.91 1.13 3.23 1.29 4.0 6.90 52 3.25 3.59 .91 4.40 1.57, 3.8 4.10 70 4.4 10.50 45 3.4 5.00 68 2.45 2.85 69 2.8 3.40 61 5.25 7.05 62 4.0 5.20 63 3.6 2.87- 1.26 3.02 1.36 4.10 4.73 0.87 5.06 2.08 3.0 3.40 0.88 3.86 1.30 2.2- 1.97 1.16 . 2.11 1.35 2.35 2.07 1.14 ' 2.46 1.38 4.55 4.37 1.04 : 5.05 1.40 3.35 3.28 1.02 3.92 1.33 % Mag EXTRACTED Q/MC 90 88 ' 79 J 82 90 82 95 94 88 93 84 * : 87 - 84 Q =- (% mag) x MC' m * MR rdg./ Q -- 68 -- UCC. 004634 MC - MR/n, _t d t. y1 f t a s r r' d 'j e e f % n t it o h f .t n c s e Table II Effect of Additional Time in Waring Blender Fiber Grade 6D 1 min................................ 15 min................................... 7R 1 min.................................. 15 min.................................. 7RF 1 min.................................. 15 min.................................. ` Total Weight MC 6.80 6.35 3.70 3.20 7.05 6.00 Magnetic Extract Chem. Anal. % Mag. MR Reading 50 3.5 60 3.4 62 2.7 69 2.65 63 6.0 71 5.15 Actual Mag.. Q 3.40 3.21 2.29 2.21 4.44 4.26 from the fiber and rock? Three samples were beaten in the Waring Blendor for 15 minutes at full speed, instead of the usual 1 min ute. The results are shown in Table II. Additional blending time does remove a significantly greater quan tity of magnetite, but not enough more so that any reasonable time is likely to make the method satis factory. Conclusions There are many Canadian fibers for which the magnetic separation yields an estimate of the magnetite content which is just about as good as the magnetic rating. However, there are a few, mostly 7 grade fi bers, which give an estimate which is far too high. There is, of course, no way of knowing these a priori, so the estimate by direct separation alone does not seem to be a likely candidate for a practical test. The extended magnetic rating method, with a comparison sample obtained by magnetic extraction and corrected by chemical analysis, of fers a way of estimating the mag netite content which is accurate in principle but is time consuming. We have no more reliable way to check it, although this might be done by fiber dispersion. T aking the extended method re sults as the standard shows that the conventional magnetic rating test is good for many fibers, but it may lead to errors of the order of 25 per cent. The error is largest for the fibers that cannot be tested by the magnetic separation method, but it is not nearly as great as for that test. The reason that the magnetic sep aration does not work for some fi- bers is suggested by a microscopic examination of the extract. The magnetite particles are found to oc cur entirely encased by serpentine rock. Anything short of a very tedious grinding operation would not free the magnetite and even then some would almost certainly be lost, going into the fiber fraction. The most promising feature of this separation technique is that it seems to offer a fairly easy wav of obtaining samples of the larger par ticles of magnetite within the fiber sample. As these are the ones which are of greatest interest, our next ef forts will be dev'oted toward anal ysis of the particle size of separated fractions. We would like to thank Messrs. W. E. Read and C. L. Williams for their assistance in the experimental portion of this work. References 1. P. O. NlCODEMUs, "The Signifi cance of Iron in Asbestos Mate rials Used for Electrical Insulat ing Purposes," ASTM Bull., Apr. 59, p. 62. 2. American Society for Testing Ma terials. Test Method D1118-57 "Magnetic Rating of Asbestos Used for Electrical Purposes." 3. M. C. Shaw, Asbestos Textile In stitute, Philadelphia, Pa., Research Fellowship Report No. 24, June 1954; No. 32, March 1955. 4. M. S. Badollet and N. W. Edcekton, "Properties of Asbestos Fiers Imported into the United States," Canadian Institute of Mining and Metallurgy Transac tions, Volume LXIII, 1960, p. 10. 'i. I I UCC 004635 i