Document RjML2VM44yzdJ73ZDx5OEOYjX

KN-77-11 "AFFI&IR" PARTICLE SIZE EFFECTS KJk.77^11 DISTRIBUTION LAST PAGE I ......................................... l E..... I, J DU PONTt'fDtET"lVN';Ei. MI -OLIUVR<1S|-i&riCOMPArN*TY-\ . NEWPORT PLANT PIGMENT COLOR RESEARCH 'REPORT TITLE: "AFFLAIR" ^ PARTICLE SIZE EFFECTS WORK DONE BY; D, P, SCHUSSLER REPORT WRITTEN BY; D, P, SCHUSSLER REPORT APPROVED BY; C, W, ANDERSON PERIOD COVERED; APRIL, 1S76. TO JULY, .1977 PREVIOUS REPORTS: KN-77-11 PROJECT CODE: 81~1Q2 NOTEBOOKS; E-837Q, E-8437, ErI2474, E^12478f E^12485, E-12937 TECHNICAL ASSISTANTS: L, I, ASAY (APRIL, 1976 TO MARCHf 1977L P, CERESINI (MARCH, .1977 TO JULY, 1977). ABSTRACT Smaller mica, particle diameter leads to a wore opaque, smoother looking l'Af flair'*, Two. different wAf.flai" samples can, be blended together to giye a product whose luster and opacity are nearly linearly dependent on the properties of the two com~ ponents. Mica particles of diameter smaller than/^70//are nearly constant in thickness. Particles larger than **1 0l/< are thicker in proportion to their diameter, The large,, thicker particles reduce the luster and opacity of ^Afflair" by effectively, diluting the product with excess mica. DUP050108587 i OBJECTIVE The purpose of this work is to determine the effect on "Afflair" properties of the mica particle size and of mica wet classification. SUMMARY AMD CONCLUSIONS o The diameter of a mica flake determines the amount of surface area to he coated with TiO~. Smaller diameter particles have more TiOn re sulting in a more opaque product. Opacity and luster are inter-related according to the equa tions TSR + 10 TDR 1200 for unclassified mica and TSR +10 TDR 1600 for classified mica. o Large diameter mica particles are thicker than small diameter particles, which have a nearly constant thickness below diameters of 70 microns. The excess thickness of the large particles does not improve its scattering ability. Large particles then are diluted with excess mica and are responsible for the lower light scattering ability of unclassified "Afflair". PATENT STATUS There are no patents or patent filings related to this report. PROGRAM No further work is planned. PUBLICATION STATUS The material in this report is not suitable for publication. SPECIAL SAFETY PRECAUTIONS None. ENVIRONMENTAL CONSIDERATIONS None. DUP050108588 -2 ~ TABLE OF CONTENTS INTRODUCTION PARTICEE DIAMETER AND OPACITY BLENDING THE EFFECT OF NET CLASSIFICATION IMPROVEMENTS IN THE NET CLASSIFICATION SYSTEM CALCULATIONS FOR NET CLASSIFICATION PAGE 3 3 5 5 II 12 i DUP050108589 T*. 3 in t r o d u c t io n Like a white pigment, it is the function of "Afflair." to scatter light, . "Afflair" has the ability to concentrate a portion of this light into the specular reflection angle. This produces a, "shininess" at the mirror angle. It will be shown in this report, that, for a given capacity to scatter light, the effective diameter of the particles determine how much light is scattered diffusely and how much is scattered at the specular angle. Our means of controlling particle size is wet classi fication using hydroclones and a centrifuge, Details of the development of our wet classification system are given in by B Cetnar and KN*-72r>8 by D, Malizia, An explana tion for the improvement in luster by wet classification of the mica is wanting. It is tempting to say that wet classification "cleans up" the mica. But, particle size analysis has shewn Virtually no foreign matter, such as clay. It is likewise tempting to say that there is a best particle size for optimum luster, Such an explanation must be accompanied by a reason why one particle size results in higher luster than another. ^Te will show that wet classification removes large, thick mica particles which effectively dilute the product with excess mica. P. ARTICLE DIAM\ETER AND O"i P-i'A> -Ci Ij*.TiY: i - Reducing the particle size of a mica increases its sur face area by exposing new particle, edges not otherwise available. Additional Ti02 is required to coat these edges, Because the additional Ti02 is not on the surface of the mica, it will scatter light diffusely, adding little additional scattering at the. spec ular angle, Luster is reduced by the additional Ti02 because the TiD2 will block the path of light that would in its absence be cleanly reflected at the mirror angle.. figure 1 is a plot of opacity, as TDR, versus particle diameter, A straight line is indicated to approximate the trend. NE^lSS-i-D is included in the plot as an exception, This code is made from partially prercalcined mica. It is probable that precaicination opens large pores in the edges of the mica. The TiOg demand, and, consequently, the opacity, are increased Without reducing the observed diameter of the flake. The particularity, regardless of opacity, is of some importance in the market place. Modern cosmetics applications favor the {Smoother product Obtained from small particle size "Afflair", Also, tests in automotive finishes show that large particle size "Afflair" reduces gloss by protruding from the finish, DUP050108590 4 ,** FIGDRE 1 r. OPACITY VERSUS DIAMETER FOR "AFFLAIR" PRODUCTS DU P050108591 r- 5 + BLENDING Blending two unlike "Afflair" lots is now used in our manufacturing process to readjust the quality of a lot to standard Wider application of blending might be used to greater advantage. It is possible that only a few codes need be kept in inventory. All others necessary for sales could be obtained by blending. Blending of two unlike "Afflair" .samples is essentially linear. The resultant luster (and Opacity) of a blend is the weighted average of the luster (and opacity) of the components. This rule holds for more than two component systems. An example Of blending application in shown in Figure 2. These blends be tween experimental 12485-28-18 and Merck's Ti-?100-FK were made on a small Scale (0.75 gms.) . Some irregularities, would be smoothed out on a larger scale. Blending with mica is non-linear in both luster and Opacity. Figure 3 is an opacity versus luster plot of experi mental 12485-28-18 and several of its blends with mica. 'The non-linearity observed is caused by the drawdown method. At the loading level used (0.75 gms/10 gms acrylic lacquer1 the luster is on a plateau so that reducing the loading by 10% de creases the luster only slightly. Similarly, diluting the sample by 10% with mica has little effect on the luster. THE EFFECT OF WET CLASSIFICATION To demonstrate the effect of wet classification, we will compare some pigments prepared from Concord mica. A sample was prepared from a portion of the unclassified mica. Another sample was prepared from a classified portion of the mica. The iuster and opacity of the samples are compared on the graph in Figure 4. Their particle size distributions are com pared on Figure 5. By removing the largest and smallest di ameter fractions from the mica, the product becomes more trans parent and more lustrous. We see that the sample from wet Classified mica is beyond the TSR + 10 TDR "iso-quality" line, indicating a greater ability to scatter light than the sample from unclassified mica. The sample prepared from unclassified mica was screened, after hydrolysis, into three fractions: -IQ^r" , +10^^-400 mesh and +400 mesh. The luster and opacity of these three samples are plotted on Figure 6. Also plotted are samples prepared from the heavy and fine fractions from wet classification, Comparison of the similar fractions, fine, middle and heavy, shows little dif ference . It therefore has little effect on quality whether the product is classified before or after hydrolysis. DUP050108592 6- - FIGURE 2 - LUSTER VERSUS OPACITY OF TWO SAMPLES AND THEIR BLENDS DUP050108593 7 FIGURE 3 - LUSTER AND OPACITY OF EXPERIMENTAL 12485-28-18 AND ITS BLEND WITH MICA DUP050108594 8- FIGURE 4 COMPARISON OF LUSTER AND OPACITY. OF AFFLAIR* FROM CLASSIFIED AND UNCLASSIFIED MICA ' DU P050108595 9* OO fO CM O m H fHa CO co s o 3 C0O0 co J O H go o 8 u fa 0 CHO CO a 1 gw; u CO tn D1 OH fa oo 1--I AND UNCIASSXFIED DUP050108596 * 10 sa 70. >1 60 h; % Ot oo Q E-f 50 40 30 60.0 800 10 00 TSR (LUSTER) 1200 FIGURE 6 - COMPARISON OF LUSTER AND OPACITY FOR CLASSIFICATION . BEFORE AND AFTER .HYDROLYSIS DUP050108597 - ii It la shown In KN^77*\L2 that the luster and opacity of a composite of ^flalr samples axe In weighted proportion to the luster and opacity of the components* on Figure 6 Is plotted the luster and opacity of the unclassified sample calculated from the components, the classified sample, class,, heavy and class,, fine, Also shown Is the luster and opacity of an actual blend the classified sample class*, .heavy and class,> .fine In proportion to the composition of the unclassified sample* Both the calculation and Blend have similar luster and opacity to the unclassified sample, Prom inspection of Figure ff, it is clear that it is inclusion of the heavy fraction in the calculation and the blend that causes the unclassified sample to have lower luster. The. heavy fractions are the only fractions, whether.classified before: or after hydrolysis that deviate sub~ stantiaily from the t bk + IQ. t d f f 16QQ ',iso*-guaiityB' line* Comparison of Figures 3 and 6 provides a clue to why the heavy fraction scatters light less efficiently that the! lighter fractions. The luster ad opacity of the heavy fractions aim similar to sample 28^18, a middle fraction, wlth/^3CL% mica added to it. The heavy fractions act like lighter fractions diluted with excess mica. The plot of thickness versus diameter in Figure 7 shows that this is actually the case. The thickness of mica particles below 70. microns is- fairly constant* Above 70. microns r the thickness increases in proportion to the di^ amcter. The larger flakes are thicker, The additional thicks ness in no way improves the ability of a flake.; to scatter light, but effectively dilutes the product. The thickness versus dia^ meter plot in Figure 7 is probably a natural consequence of the grinding process.* There is apparently a minimum in thicks ness that is reached before a minimum in diameter. As the flakes become thinner, they become more fragile and their diameters are more easily reduced* IMPRQyBMBHTS IK THE WET X^g-glFICATIW SYSTEM Prior to this work, the wet classification system (as described in KN-68*-g.i. suffered from two disadvantages* it re~ quired constant operator attention to. prevent .the hold drums from overfilling and it was difficult to adjust the overflow to underflow ratios of the hydroclones * Much of the design of this wet classification system was dictated by the speed of the centrifuge* The. centrifuge runs at a feed rate of about 7 GPM, whereas the hydroclones run at about 21 GPU, The hydroclone would not perform effectively at lower flow rates, It was then necessary to recycle two^thirds Of the output from the hydroclones to allow the centrifuge to keep pace. DUP050108598 12 FIGURE 7 DIAMETER (MICRONS1 THICKNESS VERSUS DIAMETER FOR FRACTIONS OF ALSIBRONZ MICA, DOT 156 DUP050108599 b - 12 * IMPROVEMENTS IN THE WET CLASSIFICATION SYSTEM (Contd.) The older system recycles two-thirds of the output from each hydroclone by means of tees located in the overflow and over flow lines of the hydroclones. Diaphragms of the proper diameter are located in each side of the tee to allow one-third of the total flow to pass to the next step and recycle two-thirds of the flow. Unless the total flow into each hold drum exactly matches that drawn from the drum by the pump/ the drum will gradually overfill or run dry. Also, in order to adjust the proper over flow to underflow ratios of the hydroclones, the flow rates of every line must be adjusted. Because the division of flow by the tees is approximate and the number of lines is large, the ratio adjustment is a difficult and tedious process. Figure 9 describes a system implemented during 4Q76. The design was made by R. D. Nelson. In this design, only! the outflow from the last hydroclone has a tee. The amount recycled, as dictated by the centrifuge flow rate, is adjusted by valves in the tee. Each hold drum has a 3" line near the top of the drum to allow flow to the preceding hold drum. The level in each drum is then maintained at a constant level regardless of the recycle rate. The water feed to drum #2 is adjusted by a bubbler valve. A baffle was installed in drum #1 to prevent mica from the feeder from flowing directly to drum #2. Without the baffle, the separation of the heavy fraction is less sharp. This- design requires operator attention to fill the Combin Feeder and to change drums on the centrifuge and underflow line. The smaller number of lines simplifies adjustment of over flow to underflow ratios. * CALCULATIONS FOR WET CLASSIFICATION Methods for calculating the particle size distribution expected from wet classification are described in KN-72-8. In this work a BASICcomputer program was written which includes two refinements on the orignal calculation method. In the newer method, the flow ratios of the hydroclones may be varied independently. Also, the amount of fines removed by the centrifuge is included in the calculation. In the equation x = (1'2C3X4) (^1x 2) (I-Y3X4)(X1X2) + (Yi Y2y 3> (1) X is the fraction of the mica in each particle size range re porting to the underflow. The terms x and y are defined as; DUP050108600 0 0) rtOM W tJ +> nJ 0 a ca 0o o 0) H r-l O 0 0 >1 4J > O (0 0 fl) & >1 m ro-nr in H rl H 0) 3 > S 3. rH Vl &OM Q H 'Tj 0 T3 i--I nfl r-I O fi o S3 D S3 morf w r-l .i--I 0tn>0 SH (0 0 O> 0M 0W 00) 00 0MW 000 -r| Vl VI t-3 Si S) 0? g0 g0 r-l rrl r-l m tw tw Vl Vl Vl 000 in w> 00 rt->l-00> nr 0 >o 00 Vt H 0ca 0m ca ca 00 Vt Vi St CM SO' f0tfO0 0 00 St &tfa W W tH DUP050108601 14 CM.CPIATIOHS FOR WET CIAgSIPTCATION (.Contd,! v - Snv 1 + SnV C2) x=1 -y C3) where Sn is the separation factor and V is the overflow/underflow ratio of the hydroclone. The subscripts in equation Cl} indicate the hydroclone as shown in Figure 10, The separation factors,$n, were determined empirically for the 3M hydroclones at 44 PSI and are shown in a plot in KN-63-9, A sample calculation from KN-68-9 is included in the Table. Data from a single run was used to determine separation factors for- the centrifuge. The feed slurry contained 4 wt, % mica, the feed rate was 7 GPM and the wash water rate was 2 GPM. The separation factor was varied by changing the speed of the centrifuge, Figure 11 is a plot of the fraction of mica re moved versus the centrifuge speed setting, It is assumed that all of the mica removed is from the -10 micron range. The separation factors are used like % in the hydroclone calculation. Figure 12 is a sample output from the computer program. In practice, we find that the best conditions for Concord micas are overflow/underflow ratios m 25 and centrifuge speed - 9. For Alsibronz mica, the best conditions are overflow/under flow 25 and centrifuge speed' = 7, DUP050108602 ^ is ^ FIGURE IQ < >. SCHEMATIC OF HYDROCLONE SYSTEM DUP050108603 16 TABI^E gaMPXE CArCPI&TXON HXPRPCLONE PERFORMANCE FORECAST Particle $ize . % Range +200 Mesh. .oai -200 +270 ,00235 -27Q +325 >01 -325 +400 , 0195 -400 +30 ,038 --30 + 20 .09 -20 +10 , 235 -10 .62 y . 0244 , 0555 .2000 .3277 .4872 ,6923 ,8545 .9394 X X Feed UF OF %=# # f .9756 1.0000 .56 .56 - .9445 .9998 2.48 2.48 - ,8000 .9853 5.76 5.68 .08 .6723 .9092 7,84 7.13 .71 ,5128 .6304 17,13 10.80 6,33 .3077 .1835 20.14 3.70 16.44 ,1455 .0289 24,34 .70 23.64 ,0606 .0042 21.27 -09 21,63 99.97# 31,14# 68.83# DUP050108604 P A C T IO N REMOVED 17 FIGURE 11 FRACTION MICA REMOVED BY CENTRIFUGE VERSUS SPEED SETTING DUP050108605 18 ** *HY0R 0CY C LO ME CLASSIFICATION*** CONDITIONS: PRESSURE =0-4 pcj MAKEUP WATER RATE MICA FEED PAfE OVERFLOW/UNDERFLOW CLONE =7 GPM * Sf 3.5 i*PM 1 = PS OVERFLOW/UNDERFLOW CLONE Z = 23 OVERFLOW/UNDERFLOW CLONE 3 = 23 OVERFLOW/UNDERFLOW CLONE CENTRIFUGE SPEED FEED-PATE TO CENTRIFUGE 4 = ?5 =6 = 7 GPM WASH WATER RATE TO CEMTRIFU6E = 2 GPM FACTIONS: FEED: PRODUCT: UNDEPFLOY/: +200 +270 +325 +400 +30 +2 0 + 10 - 10 4.23 10.23 8.22 7.62 14.73 13.64 18,1 23.17 1.3B094E-4 4.21B15E-3 ' 0.243442 1,.40934: . il.094 22.7335 35.7059 28*662 % YIELD PRODUCT = 49.1064 % YIELD UNDERFLOW- = 41.8940 % YIELD FINES = 8.9937 10.1443 24.4611 19.3 328 16,5364 22 .1972 5.85236 1,24571 0.230137. FIGURE 12 r, SAMPLE CALCULATION FROM THE PROGRAM '`CLONE."' DUP050108606 Ll:?l Cl OFF t : &n *Wl -L-" 1-A- TT PVT'1-. / r TC- PA! A w r H Cm ! 1 * r a A(.:r Tr .`f 2 i. pc*; ` Tf t j. >; T f.ijrft- t y * 'r t. 1 Hfr >^r.-pc; a -, .! r~ P ' > f p v ' - T ! ppr 1 F F! "p h r in t t c a -A PiijM1 vp T is^ i T VH Tf !T F'ri T' L ' 0 HR INT H AT T Pr C. CFrP- JFU'rr cPr Tvpp r rv ! rr ? t ,,- 5 .. ' A; AL Yr J` * m =, n~\ T/ ,4-2*'. :. i- l TO PIP- FfP.A ITi JrK'T fli) Fr?A f{*! F<ai F.{;^ . F (A) .C(*M .`"I'M 1!V1 PIL L".n Fr/!- Isi Tr i;,'i tc U -i^'t ?-'t t` r A-P fi ./ r *, f-f'yr r Hi. I r T nr tf J c- * rtf-' p -i -m p >. p I r Ttj r *' (! 2 - n HH lb 1 . C30 HHjfT i,r:F PAP jrc>jT f 2FP PI f F {a, p. PA-i TvT f - v ; . r. \ : , Vlr s i a .a) 2 p. * p p p ?r * TC t- 2nri FOP v!rl TC S 1. FT P ; ;,!i 51 m t FT v f I . .J1 r F {I.\ / (i .c ; T . ) i y,r>-rt FT V f I n "1 -Y ( T * ' 0' t'C'sf'T ,; r&'XT j TF'f \f:T ric'1 *r,PSK PF 1 3-n FUf- -I 0, TO T V'S !V'P >i 'T: :a A''; DIP 1.FT i FT l >*t I FT XI (IV Aft.( i - Pf ! 3 ..n ' '/ CAr.{ Y * 1 . ,.n *y ( xi (,n = a ./(; 15 (,.' 5 ~.v i (J\ feF (t i * f V ' - . ,.'l ! 0 ..T *y (<+.. M HfX T s! l.FT i 2:"p PIH U \ n 5 rr-v J r 1 TV P, 1 c*y U'jV-Ff J\ --P A-fT ii ! *r~ r~ ^ ^ V! * l \ ' ~ ' ( L. 5 T- f \ <-r f a \ v-r ur- f p s .- FT !.FT Y'O'-' !L. i? Ji tj \ T.\ ' (A J i isu; V) -h 1 (2 '-v ' ") . i-- i T r p (#i > h i C(,!T!r(J! XT2) *1^1. hV^l F;-') KT'M DUP050108607 p 'i n f -FV 1 , S 'in FOP J * i- tc r SAA t FTUt. ~/1 ?fji ) -*fiO >. F **!} f ;Ky r ,.! S A A ft'IM pp A PP JfT ft ' i ' h-plry ,AP (-51 ) i ** > l 'f'DPOri.f'f'F rl AF' If 'rAT? Of '-fe ** ~ '* 61A pPIfT ft?A. HP Tf-'T ft~0 TA^c>p AP ftuo h PJ.p T ftp A Wit'f o~- A p p Iff r At"11 'rt rtri i icr?: p p f F''u c r* ft "'A FFlri TAP-{ ?.P*. *, * i-'A ' FtJ P 4' A T.Fft P A1' F s p GPP" 680 KP If M TAP (?P1 5"vIf'A FFrD HATF S 3.P FPF 6nn rrif'T TAP{API :-*OvFFF-l..0'*V1 'NOFF-Fl CW Cl. CNF 1 s ;p (i i FA * Kpirr :i`AP( API r " OvFFFt.Ctf/UtTFFFI QV Cl 0?ip t. ?'? ft { ^ i - 7 m PR If.`T TAC { APT ? ''OVFHFt 04/UNOFFFI O'V CL Ct'F P 7 pp T:f 'T -AP(PP) *. '* Cv FHFLO'l'/UrTFPn 04 . PL CNF 4 `Ha p.P pT. ! jM ( P P) ? '*CFf' I F IFur-F SPF1 P r"fP(Pv S'*.? P < 4 i s'*;C F'Pt a 'T 1/ip (p p i ' ''FFf p PAtr t o cfrTFjFur-F OFF** HPIPT iAPI'^lV'vAVH I'AtFP i' A f F 10 Or? TFjFUGF 2 ^ ppr * r- 4`V P-SJ" l - ir? ' t-'P Tf !T PPI.--7 ! r ? ^ 5 (r.P-A f .1 lor1 rn P HFCOUCT l.*prFFFi POINT f O 1 s' 7 ( 1.81 - '~^o > ' * TAP {pPl *F(11* TAF'i p^T *U(1V * f`-(i. C-f rrT ra?( *>A1 ! * *''*' "< ' * .TA'P.{?*>' FA PK?T ..' "A 1 U { P11 -T*F t lifti * P {s p.c Tf-f TA'F.{.1?11 ? pp TpT TAP- r \ r. \ ,'TAP('^1 `FT si *FAPT*1ftV;U{~1 iTAPfiiM !pn ! yt.pi f 'sn >v5-F( u v* T 'P ("Ai p j (iai \?t rTA*. pp /{ff.>-A i? fC'< a pp jp.= T T AP ( ' + - A !* ?. r,Ap f FP1 * F( F } r r* (Pp ) !Uf p i fTAPj?A*> rp-<r?:i ^ pji'T AP ( IP'S i.-'n *: I'Amosi ip{a t ' ! tA-i i U (A 1 : T AP f a*. 1 *- P > A1 r, i_v t t- P I NT r a p . (i. ' i x \ T * i TAP ( l : F {? \ T a r*! AR1 ? L';" > TA1'! tA > p r - ) I PHU-T T A P (i - 1 * A * * T ftf? ( OPT T *.r=; m.i/o ) T A P (uV1 *r ' " } I PPJNT i PP.S-VT JAB($ i) ; > ;< f t f i. o PPrrtjrr a -* t i FPlrT TAP(pal *< Y I Fl o Of'T'FPf'!. C s'"' , PPICT iApfTA! M < YIFLP F-INFS i Pt-'INT i PPIPT I PRINT i FP IPT ^At T A Pi. OTP'H i .7 f 'f~VT po; - IF R^yr-YFP;* TFFF ^"1 i CC Tc 1PA'- i Ffp Irr ITT- i < X"S{ n=U( T) IF TA'f 15 v-iet Ti Ft' 1111 F '/ "> ( I Vr T , , f'FvT I v"i Pr Irr f A { "i) '"rT . A' n p p I r T I A f( ''I 5 * n 1 Ar ; 1 -* \ * - t\ > f T AF- ( ACrn ? i i *r: 1 I u" \ ' * ' . > *^p If'T I AP t "1 ""'---_. ~ - ,-- ...... ` i pPj f'T *n,F,cm\ * * - f If-T T.AP / - } T t f ' i Jt -j T -Ap r*)S * < If'T '* -1 *( i *; 1 -i-If'T LFFT-if-'X/ .- .,v .. ; v DU P050108608 l *2-i ppjt':'r _ap i * u lA "VI PH I FT t AP ( T) ? > * * r.w PHJFT ,T AP {-) ? *' * ' *3'I h p ip t '* +i/> ! * Afi h p in f FFT M ? -% . 'i. < * >>! HP 1! T ( Ar' (7 1 ' * ; :. '.'I yr-yij 1 ft.p f -- 1 * * 1 JV* ;'| pi " i - r l a !? ( -- ' * i.'in m f-'T * p- -i- 7 :f f `i ! "3 " r< h-T l,Ff- I T ( "V ' v: i' L ? ..1 3 A. >. T "T t-'pjrr t wT( r pp tf -y HP If T r,P '' s 1 p I ' / ** 1 t 1 * ''A'* * TAR ? " J w '' 'T'! I AR ^. ] t~!i >*? l?3n ! PT { T t FF t T (''y. Y. / . '* mi HPlr'T *t *: TAP ( f v* ' t * i *r -< HPjrT ipi y ap*. ( -- ] f !' * l ^3 i p p In . '; TAR ( - \ : * * * 1 a t > 1 PJfT i ~ * *j ppjr-7 l Ffl * ( 'X/ " &i - i"n 1 I-."! HP I> T HPlf'T v.*.p f "i * i /\p ^ j i i Tm HP If'T i a c' { ^\ : ' 1 "3 *1 fp. I py { .It- PP Tf T X FFiTC'X./-'.; t y \ S' i lv*o p p jr 'T j r R { 7 \ . -- t 1 tspr, HPTt't i AB {->' * pp Te-' T I AB f - 3 * " * m HPlf-'T if [-( l-PIf'T LFFi T C' (^' I U -! ", pp ip r I'AB *- t *>"q pplri TAR (-1 V4*' * itff', PPJHT ' a p {n ; - 1 ' 1 us : ) Hpj FT -* fpT 1 't I'-AT HP JFT l FFT a?*'- / * (in 1U7I1 p p i rr TAP ("l * * ' ' HP I ft: T i a p i ; " * ' 1 U'7 T ppirT I AF.r* \ f' * J` 13rV. Hp7ftfT -AI'DTIFh rUHfP'M I '' HUT ft4*. IS 20 IF f s --,ypC' Jj .pf: ' ssrpl lS"Vf pp Tft'T ''c h a '-h f rcf-r.ir?o ISU 1 If Vi'T KT 1S--T if y~ r** f FF'' Tt-Pf i ITT Ir;T- f'AT A. \ f 1 2^. U is- n FtT 1 vn \T * n .-n DUP050108609 E , JC^nikkN^filmington &,W, S,\stW*rK Newark. E, Ne<3 *> Exppeerimmeenntal Station 4, j, E, Romano ^'Newport 6,5, 6, A, Hapka ~ Legal ~ Wilmington E. E, tfaffe/Circulate/File < * Newport 7, c, w Anderson ~ Newport 8., ; R, Z, Fortnay/V, A. Roraito ~ Newport 9., J, C, Chaney Newport la. A* X>,. Volk - Newport 11, :R, J>, Nelson Newark 12, D, F, gchussler w chestnut Run 13, Central Report index -r chestnut Run 15, Newark Research File r Newark IS, Numerical File *> Newark 218 17, Library * Newport 18, Extra 1SL, 2a,. DUP050108610