July 1949
INDUSTRIAL AND E N G I N E E R I N G CHEMISTRY
pensive dispensers have now been made available for insecticidal aerosols, the liquefied gas method will be expected to expand into many fields where an aerosol or a wet spray is useful. LITERATURE CITED
(1) Dietz, W., and Shepherd, H. H., Soap Sanit. Chemicals, 22, (12) 137-9 (1946). (2) Fulton, R. A., IND. ENG.CHEM., 40, 699-700 (1948). (3) Goodhue, L. D., Chem. Inds., 54, 673-5 (1944). (4) Goodhue, L. D., IND. ENa. CHEM., 34, 1456-59 (1942).
(5) Goodhue, L. D., Fales, J. H., and McGovran, E. R., Soap Sunit. Chemicals, 21, (4) 123, 125, 127 (1945). (6) Goodhue, L. D., and Hazen, A. C., IND.ENG.CHEM.,ANAL. ED., 19,248-50 (1947). (7) Goodhue, L. D., Schultr, F. S.,Innes, Neva, and Stansbury, Roy, Soap Sanit. Chemicals, 23, (9) 119-21 (1947). (8) Goodhue, L. D., Schultz, F. S., and Wilkins, P. H., Chem. Inds., 60, 602-4 (1947). (9) Hazen, A. C., and Goodhue, L. D., Soap Sanit. Chemicals, 22, (8) 151, 153, 155 (1946).
1527
(10) Ingle, Lester, J . Econ. Entomol., 40, 264-8 (1947). (11) Jones, G . W., and Scott, F. E., U. S. Bur. Mines, Rept. Invest. 3908 (1946). (12) Madden, A. H., Schroeder, H. O., Knipling, E. F., and Lindquist, A. W., J. Econ. Entomol., 39,620-3 (1946). (13) Neal, P. A., yon Oettingen, W. F., Dunn, R. C., and Sharpless, N.E., Suppl. Pub.Health Rept. No. 183 (1945). (14) Rhodes, W. W., and Goodhue, L. D., Soap Sanit. Chemicals, 23, (10) 122-3, 151 (1947). (15) Rotheim, Erik, U. S. Patent 2,128,433 (Aug. 30,1938). (16) Smith, C. M., and Goodhue, L. D., IND. ENG.CHEM.,ANAL.ED., 16,355-7 (1944). (17) Smith, W. W., Baldwin, Yvonne, and Grenan, Marie, J. I n d . HDg. Tozicol., 29, 185-9 (1947). (18) Stoddard, R. B., Soap Sanit. Chemicals, 24, (7) 147, 149 (1948). (19) Young, E. G., Ibid., 23, (11) 116-17, 152A (1947). RECEIVED June 11, 1948. Presented as part of the Symposium on Aerosols before a joint session of the Divisions of Physical and Inorganio Chemistry and Colloid Chemistry a t the 113th Meeting of the AMERICAN CRE:X.CAL SOCIETY, Chicago, Ill.
Composition of a Synthetic
Gasoline ALFRED CLARK, ANTHONY ANDREWS,
AND
HAROLD W. FLEMING
Phillips P e t r o l e u m Company, Burtlesvills, Oklu. T h e percolation through silica gel of eight fractions from a gasoline synthesized from carbon monoxide and hydrogen with a fluidized iron catalyst of the synthetic ammonia type has accomplished the breakdown of each fraction into the following four types: paraffins, olefins, aromatics, and oxygenated compounds. As a result of the segregation of these types, the following additional information was obtained on the character of the gasoline fractions: (1) Forty to fifty per cent by volume of the Cg, Ce, and C7 fractions consist of straight chain 1-olefins. (2) Fifteen different aromatic compounds were identified by infrared absorption as being present in the first six gasoline fractions. (3) There is fairly conclusive evidence of the presence of small amounts of diolefins in most of the fractions.
T
HE efficacy of silica gel in separating mixtures containing different types of hydrocarbons is now well known. Mixtures containing paraffins, naphthenes, olefins, and aromatics may be separated by percolation through silica gel. Several analytical methods using silica gel have been developed for determining the, volume of the various types in hydrocarbon mixtures. Mair (3) outlined a procedure for determining the aroinatic content of a straight-run petroleum distillate, as in the gasoline or kerosene boiling range. Lipkin et al. ( 2 ) describod a method for the determination of aromatics in petroleum fractions boiling above 400" F. A method for the analysis of small samples of shale oil naphthas was developed by Dinneen and eo-workers (1). The present report describes in part the results of large-scale laboratory silica gel percolations in the analysis of eight fractions of a synthetic gasoline prepared by reacting carbon monoxide and hydrogen in the presence of a fluidized iron catalyst of the synthetic ammonia type a t approximately 600" F. I n addition, reaults are given for the precision fractionation of some of the segregated percolates. Since this paper is concerned with separa&ionson one representative synthetic gasoline, no discussion or conclusions concerning other types of gasolines are intended.
MATERIALS AND METHODS
MATERIALS.The synthetic gasoline was first debutanised and then separated into eight fractions by an Oldershaw glass bubbleplate column. Table I summarizes the boiling ranges of the eight fractions and the chemical tests carried out on each fraction. TABLE I. PROPERTIES OF SYNTHETIC GASOLINE FRACTIONS Bromine Hydroxyl No. No. c6 59-104 80.6 183 10.7 C6 104-165 80.8 154 14.5 c7 165-219 81.3 133 21.8 cd 219-279 78.4 112 17.9 Ce 279-324 79.5 101 15.3 ClO 324-369 81.4 93 12.0 Cti 369-394 84.7 88 9.9 c 1 2 394-428 85.0 81 9.3 a As determined by bromide-bromate method.
Fraction
Boiling Range,
F.
Olefin",
%
Carbonyl No.
Acid NO.
2i:o 20.4 23.8 14.7 7.9 5.3 9.1
1.1 2.8 2.5 2.6 1.7 0.7 2.6
...
The silica gel, purchased from the Davison Chemical Corporation, graded as follows when screened by a Roto-Tap: Mesh Size 150 150-200 200-250
Wt. % 19 23 8
Mesh Size 250-325 > 325
wt. % 20
30
Regeneration of the silica gel was not attempted; a new batch of silica gel was used for each percolation. Absolute ethyl alcohol was the desorbing liquid in each of the percolations. CHEMICAL TESTS. Whenever needed, the following group of chemical tests was carried out: Hydroxyl number determination involves the esterification of the hydroxyl group by a 1 to 4 acetic anhydride-pyridine mixture. The excess acetic anhydride and the acetic acid resulting from the esterification a m titrated by a standard potassium hydroxide solution. The hydroxyl number is calculated as the number of milligrams of potassium hydroxide required for a 1-gram sample. The hydroxyl number must be corrected for any acids present in the sample.
INDUSTRIAL AND ENGINEERING CHEMISTRY
1528
Figure 1. Stainless Steel Adsorption Column with Jacket
--
B
A. B.
C. 11.
E.
F. G. H. 1.
Brazed 1-inch pipe nipple Stainless steel tubing, 3 feet long and 3-incli outside diameter Stainless steel tubing, 3 fect long and 5-inch outside diameter Jacket of galvanized sheet iron Stainless steel tubing, 3 feet 4 inches long a n d 1.55-inch outside diameter Stainless steel tubink, 3 feet 4 inches long and 0.75-in011 outside diameter Removable micro mctallic plate Brass collar h a z e d t o t u b i n g Brass dripper threaded to collar
Vol. 41, No. 7
After all the gasoline fraction had entered the silica gel, ethyl alcohol charging was begun. Piitrogen pressure was slowly xpplied to the alcohol, and was so regulated that the gasoline fraction, having been charged early in thc morning of thc first dag', r ~ o u l dappear a t the bottom of the column early in the morning of the second day. From previous experiments it was possible t o calculate the approximat,e pound-hours required for a fraction to percolate through the column. In most of the percolations the flow rate of the filtrate down the column averaged about 15 em. per hour. The rate of collection of filtrate varied from minima of 20 nil. tjo maxima of 60 ml. pcr hour.
TABLE 11. DIAIBXSIOSS O F lJERCOL,4TIOS Column Material Volume of rescrvoir. mi. Silica gel ieotion, e m . T o p section, i.d. TOFsection, length Second section, i.d. Second section, length Third section, i.d. Third section, length B o t t o m section. i.d. Bottom acction, length Volume of oolnmn, nil. Silica gel, grams
Total acid number is determined by d.S.T.31. D 663-46T. Bromine number is determined by A.S.T.i\I. method D 875-46T. Carbonyl number method involves the reaction between hydroxylamine hvdrochloride and the carbonyl group of ketones and aldehydes. The liberat,ed hydrochloric acid is titrated with standard alliali. The carbonyl c nuinber is calculated as the number of milligrams of potassiunl hydroxide required for a 1-gram sample. The carbonyl number must be corrected for ariv acids uresent, in the sample. Alaleic anhydride method involbes the reaction between a conjugated diolehn and a solution of .5yomaleic anhydrlde in toluene. The unreacted maleic anhydride is ,separated and t~tratedwlth standard alkali. The maleic anhydrlde value is calculated on the basis of 2 moles of iodine reacting wit,h each [no!, of diolefin, and is defined as the number of grams of lodme ctyuivalent to the olefin double bonds in 100 grams of sample.
Glass 3000
COLUMSS
Netal
....
5 0 88
7.3
4 0 88
1.7
2 5 100
1 0 101
7100 2230
88 88
2.5 100 1.3 100 5300
4410
T o minimize any possible polymerization effects and to prevent loss of volatile cornponcnts, ice water a-as kept circulating through the jacket of the column during the percolations. Highly volatile fractions, like the Cj's, were charged and the filtrates caught iTere iced down. Charges of less volatile fractions \ cooled t o a filtrate temperature of 68" F. As each filtrate fraction was collected and its volume noted, the refractive indcs was determined. At the conclusion of percolation, a graph was plotted of the volume of each filtrate fraction against refractivc: index. From this graph were calculated the respect,ive volumes of paraffin, olefin, and aromatic. The cut point between paraffins arid olefins, olefins and aromatics, and aromatics arid oxygen compounds mas taken as the centctr of the most vertical portion of the curve. Whon there was doubt, bromine numbers were determined on the critical fractions hetmxn paraffins aiicl olefins and olefins and aromatics. The cut between aromatics and oxygen compounds was checked by testing for cvolutiori ol
PERCOLATION Co~carxs. The dimensions of the tn.o percolation columns are given in Table 11. The glass column ab used in thc percolaI I I tion of the Cg and the CTi $ractloilq The metal ' I I I (stainless steel) column (Figure 1j tvaq used for the remaining six fractions. These columns, \\hose design was the result of previous work, were tested with known mixtures of 1-oct.ene and n-heptane before being used in the percolation of the synthetic gasoline fractions. Both columris give sharp separation of n-heptane and 1-octene. The volume of filtrate intermediate between pure n-heptane and pure 1-octene was 1.1 mi. when a mixture containing 225 ml. of 1octene and 275 i d . of n-heptane was pc3rcolated through the glass column. The volume of filtrate intermediate bet,vieen pur(: ti-heptuno and pure 1-octene was 2.4 mi. when a mixture coniaining 200 inl. of n-heptane aiid 500 ml. of 1-octenc was percolated t.hrough tho steel I I -_L L column. 1.3600 1 The experimental procedure \vas standardized 1 1 as much as possible. A4fler the column :,was , i I packed with t,he desired amount of silica' gel, loo 150 200 280 300 350 400 50 thc gasoline fritction was permitted to percolate FILTRATE VOLUME- ML, into the silica gel bY gravity and bY the applicaFigure 2. Results of Percolation with Silica Gel of Ca Fraction (Boiling t , o n of a pound or two of nitrogoii pressure. Range 59" to 104" F.)
390 1
~ t -
I
j
1
INDUSTRIAL AND ENGINEERING CHEMISTRY
July 1943
1529
1.400C
NJ
OF
MIXTURE
1.39OC
~
N
'D"
COMPOSITES SEPARATELY
A
aB
FRA~TIONA+ED
1.380C F E E D ' O L E F I N P L A T E A U MATERIAL FROM PERCOLATION OF Ce F CTlON
1.370C
IJ60(
. FILTRATE VOLUME-ML.
Figure 4.
20
1 40
60
Results of Repercolation with Silica Gel of Cs Plateau Olefins
Figure 3. Results of Percolation with Silica Gel of CB Fraction (Boiling Range 104" to 165" F.)
hydrogen with metallic sodium. The ethyl alcohol effectively diluted any high index oxygen compounds so that a sharp break was obtained in refractive index. I n determining the volume of oxygenated compounds, a recovery of 100% was assumed. The volume of oxygenated compounds was taken as the difference between the volume charged and the total recovered volumes of paraffins, olefins, and aromatics. Therefore, in case any hydrocarbons were lost either because of high volatility or by being left behind in the column, the volume of oxygenated compounds would be too high. This method of determining the oxygenated compounds was necessitated by the fact t h a t the ethyl alcohol did not completely displace the oxygenated compounds without a t the same time mixing with them. DISCUSSION F I L T R A T E VOLUME-ML.
T h e results of the silica gel percolation of the eight synthetic gasoline fractions are plotted on Figurcs 2 to 10 and summarized in Table 111, which indicates some definite trends. The paraffin content of the higher boiling gasoline fractions is lower than that of the lower boiling. The paraffin content of 1.8% by volume for the Cg fraction appears to be out of line, probably because some of the Cg paraffins may have been lost during the run. The olefin content of the various fractions fluctuates between
TABLE111. RESULTSOF SILICAGEL PERCOLATIONS, IN PER CENTB Y VOLUXE
Fraction
C6 CP
c7 CS
CS ClO CII Cl2
b
Paraffins 15.4 13.2 10.5 9.6 1.8 6.0 9.3 5.4
Olefinsa
Aromatics 0.0 0.2 2.1 6.0 6.2 7.2 5.8 3.6
Oxygenatedb 2.8 6.6 7.2 8.1 13.0 7.5 10.9 14.7
Figures in parentheses were obtained by t h e bromide-bromate method. By difference.
Figure 5. Results of Percolation with Silica Gel of Fraction (Boiling Range 165" to 219" F.)
C7
74 and 82% by volume. Except for the C,, and Clr fractions. the agreement between the results of the two methods is close. The values obtained for the aromatic content fall in line, with the maximum concentration being 7.2% by volumc in the Clo fraction Since the oxygenated compounds are calculated by diffei encc, any errors in any one of the other three types would cause th s value to be in error also. T h e high value of 13.0y0by volume for thc oxygenated content of the CS fraction probably results from errors in measuring one or tmo of the other types. In t h a t case the percentage by volume of the oxygenated compounds in the first six fractions varies between 3 and 870,with an apparent increase above the maximum figure for the Cil and the C1*fractions. Except for the Cb fraction, additional analytical work was carried out on the percolated gasoline fractions. A low temperature distillation combined with a selective absorption of pentenes
8 ANALYZED
BY I N F R A
1.3700
erized over the silica gel used in these experiments a t 0 " C. (32" F.). Therefore it was assumed that no isomerization of mono-olefins was
RED
1.3600
I
LLL
-
-
obtained on a plateau of refractive index (ng) 1.4090, and 2 to 3y0 was higher-index octenes. Upon repercolation of this plateau material, 98 to 99% was recovered as plateau material of refractive index ( n g ) 1.4090. No high indcx octcnes were obtained, which establishes t h a t none of the 1-octene was isomerized. Phillips research-grade 1-hexene (n%' 1.3880) was percolated through silica gel a t 0" C. The recovery n : ' 1.3880) was 98 t o 99%. of I-hexene plateau material ( The results in Table I V indicate that the hexene isomers may be separated by silica gel percolation with a column of sufficient height, and that heptenes are less strongly adsorbed than pentenes. The order of increasing adsorbability of the hexenes on silica gel for this particular c g fraction was: 4-methyl-l-pentene, 3-methyl-l-pentene, 1-hexene, 2-hesene, 3-hexene, and NOTE
1.3900 1.3800
I
1.3700.
1
A AROMATICS F R A C T I O N A T E D 8 ANALYZED 01 I N F R A RED
i
cFigure 8.
Results of Percolation
I N D U S T R I A L - A N D E N G I N E E R I N G CHEMISTRY
July 1949
3-methyl-2-penteLe and 4- m e t h y 1- 2- p e n t e n e . This order may be changed with other mixtures, depending upon the concentrations of the individual olefins. A 100-plate fractionation of the CO paraffins indicated t h a t 80% was n-hexane, 10% 2-methvlpentane, and 10% 3-methylpentane. T h e composition of the CP, fraction, in volume per cent, exS C7's, follows: cluding C ~ and n-Hexane 2-Met hylpentane 3-Methylpentane 1-Hexene 4-Methyl- a n d 3-methyl-1-pentene High index hexenes Benzene Oxygenated
10 1 1 49
6 3 3
6 5.6 24 8 0.2
1531
1.5100
1.5000 I4900 14800 1.4700 1.4600 I4500 1.4400 1.4300
'
I4200
I4100
6 6 -
1.4000
100.0
1.3900
i/
I n a n effort t o determine the composition of I.3SOQ the C7 olefins, three olefin composites (Figure 5 ) 1.3700 were fractionated separately (100-plate column). I3600 Composite A, containing 196 ml. of the total 0 30 80 90 120 150 180 210 240 270 300 330 FILTRATE VOLUME- ML. predomi281 ml. of C, olefins, appeared t o be nantly l-hePtene. Fractions on the low side Figure 9. Results of Percolation with Silica Gel of Cl1 Fraction (Boiling Range 369" to 394' F.) (boiling of 1-heptene were contaminated - point) b y a compound like 2-methyl-1-hexene; on the high side compounds like 2-methyl2-hexene and 3-methyl-2-hexene seemed t o interfere. About 13% b y volume of composite A boiled in the range of the 3-, 4-, and 5-methyl-1-hexenes. T h e remaining 12% b y volume consisted of high boiling heptenes and low boiling octenes. About 60y0 by volume of composite B, which contained 38 ml., consisted of material boiling near 1-heptene. About 10% by volume boiled definitely below the 1-heptene range, and the remaining 20% consisted of high boiling C? isomers. T h e fractionation of the high index material in composite C was not of much assistance in the identification of its components. * The fractions in composite C ranged in color from light yellow t o deep red. On the basis of refractive' 0 index, the kettle residue from the fractionaFILTRATE V O L U M E - M L . tion of composite C had undergone some Figure 10. Results of Percolation with Silica Gel of C12 Fraction (Boiling polymerization, an indication of the presRange 394' to 428' F.) ence of highly unstable or reactive components such as diolefins. T h e aromatic-containing filtrate fractions from each percolaCi2 fractions were not determined because sufficient data on tion were composited and, in turn, fractionated in a 100-plate known compounds in that range were not available. column. Infrared scannings were carried out on the fractionated Although i t was not possible t o isolate and identify any pararomatic cuts. The following aromatic compounds were found: ticular diolefin, the presence of small amounts of diolefins in benzene, toluene, o-xylene, p-xylene, m-xylene, ethylbenzene, most of the gyoline fractions is indicated by some of the data. n-propylbenzene, isopropylbenzene, l-methyl-2-ethylbenzene, The percolation graphs show the presence of high refractivc index l-methgl-3-ethylbenzene, l-methyl-4-ethylbenzene, n-butylbenzene, isobutylbenzene, l,a-diethylbenzene, 1,4-diethylbenzene. T h e possible presence of the following was also indicated: 01TABLE V. DATA INDICATIKG DIOLEFINS methyl styrene, isobutylbenzene, l-methyl-4-isopropylbenzene, Fraction or Composite Fig. No. of Cumulative Unsatd., Malelc anhydride and tert-butylbenzene. The aromatics present in the Cll and No. C atoms vol. n 70a value
-
10 10 10 10 8 10 8 10 9 11 9 11 9 11 9 11 10 12 10 12 10 12 Determined bromine number 8
TABLEIV. Fig.
No. 3
4 4 3 3 3
SUMMARY OF FRACTIONATIONS OF Cs OLEFINS,IN VOLUME PERCENT
Olefin Composite
D A B
A
B C
1-Hexene 44 73 78 48 0 0
3- and 4-Methyl1-pentene 24 10 3 4 0 0
High Index Hexenes
0
0
19
48
94
76
Pentenes 0 0 0 0
6 24
Heptenes 32 17 0 0
0 0
8 8 8
408-458 1 4259-1.4266 100 480 1.4292 105 490 1.4333 114 495 1.4380 120 500 1,4442 129 503-512 1.4570-1.4839 122 1.4312 102 22 1 23 1 1.4350 110 235-240 1 4400-1.4468 124 245-250 1.4362-1.4830 115 235 1.4331 102 240 1 4345-1.4473 114 258 1.4498-1.4834 140 from bromine number of fractions; for Cio, Cii, a n d Ciz mono-olefins.
.. ..
.. ..
1.i
1 11 assumed theoretieal
1532
INDUSTRIAL AND ENGINEERING CHEMISTRY
material on the tail end of the olefin plateaus. I t would be expected that the diolefins, being more strongly adsorbed than the olefins but less than the aromatics, would appear at the tail end of the olefin plateau. T h e presence of aromatics having a lower viscosity than the olefins is not a satisfactory explanation iri this case, since close-boiling fractions were uaed in all the percolations. Determinations of olefin content on the fractions at the tail end Of the Olefin plateau by the hrolnide-bromate method have given several anSWers substantialllv about 100% and maleic anhydride values as high as 17. The data are sumniarized in Table V.
Vol. 41, No. 7
Diolefins are indicated by bromine numbers as high as 140. Conjugated diolefins are indicated b y maleic anhydride valueranging from 11 t o 17. LITERATURE CITED (1) Dinneen, Bailey, Smith, and Ball, IXD. END.CHEM., AXAL.ED.. 19, 992 (1947). (2) Lipkin, Hoffecker, &fartin,and Ledley, Z&j., 20,130 (1948). (3) Mair, B.J.,J.Research Xatl. Bur. Standards, 34, 435 (19451. RECEIVED April 10, 1948.
ANTIFOULING PAINTS Studies in Multiple Pigmentation ALLEN L. ALEXANDER AND R. L, BENEMELIS' .VacaE Research Laboratory, Washington 20, D . C .
T h e wartime shortage of cuprous oxide inspired many investigations of substitutes and extenders for this primdry ingredient of so many successful antifouling paints. A s a result, metallic copper flake and powder were used to a degree hitherto considered unsafe on steel hulls for fear of accelerating galvanic corrosion. I n the course of the work reported here, a choice had to be made in the selection of inert pigments to be used with specific toxic ingredients. Zinc oxide appears to provide considerable advantage over more orthodox inerts such as diatomaceous silica for use with cuprous oxide. Similar differences exist but are somewhat less pronounced with respect to metallic copper pigments, while some mercury pigments appear to function equally well with a wide variety of extenders. Experience with metallic extenders, iron and zinc powders, is described with special emphasis on their limitations.
T
HE first article in t,his series ( I ) demonstrated that heavy
pigmentation with metallic copper, in the form of flake or powder, could result in complete inactivation of the protective qualities of antifouling paint when applied t o steel. Some additional evidence is presented here to substantiate this view, Young and others ( I O ) studied the effect of extending antifouling pigments with a n inert material (barytes) and advanced a theory as t o the probable role of such an inert in promot,ing the efficiency of the primary toxic pigment. During the recent war demand for the widely used cuprous oxide greatly exceeded the supply, and resort was made t o other copper pigment's whose ability t o suppress fouling attack had been adequately demonstrated (9). I n each of these studies the necessity for close control of pigment-volume relations has been emphasized in order t o avoid serious consequences such as would result from local galvanic couples, a rapid depletion of the store of available toxic, and the complete inactivation of the antifouling film. T h e Woods Hole, Mass., investigators made a comprehensive study ( 3 ,4, 6, 7 )of the factors governing the action of antifouling paints and showed that matrix and pigment solubilities are critical in the maintenance of adequate leaching rates (8). Leaching rates have been more or less accepted as an accurate measure of ability t o prevent the attachment of fouling organisms. Young, Schneider, and Seagren ( I O ) demonstrated that t h e substitut,ion of nontoxic pigment for an equal weight of vehicle improved the ant,ifouling qualities of a copper paint. This increase was at,tributed to an increase in the permeability of the film. Thus, if pigment volume is a factor in permeability, its 1
Present address. Sinclair and Valentine Company, Ridgeway, Pa.
regulation between established limits, whether by the addition of toxic or nontoxic pigment, beconies a significant factor in ant,ifouling effectiveness. The second article in this series ( 2 ) .showed that certain mercury paints containing low volumes of pure toxic are considerably less effective than more highly pigmented paint's containing cven smaller amounts of toxic but increased amounts of inert pigment. Ketchum and Ayers (6') indicated that some differences exist between inert pigments used with certain toxics. This evidence points to the need for determining the more exact role of inert pigments in antifouling paint formulat,ion. Materials considered for this application include, in addition to the usual inerts, zinc oxide, zinc dust, iron pori-der, and many of the prime pigments used in orthodox paint formulations. Primary pigments are referred t,o here as those possessing antifouling or toxic value where hiding qualities are of no consequence. Aside from the solubility characterjstics of toxic pigments, rate of solution is governed b y the extent to which they come in contact with sea water. This, in turn, may be a function of the solubility of the matrix, the rate at which it exfoliates (underwater chalking), its porosity, or some combinat,ion of these factors. I n these experiments att,cmpts werc inad(, 60 study the effect of adding graded amounts of inert pigment on t,he ant,ifouling properties of the paint. An endeavor was also made to accelerate exfoliat'ion rate through the addition of corrodible iron and ziiic powders. In other esperiment,s t\vo t,osic pigmenrs were studied in combination. In each experiment direct comparison was made with standard pigmentations for control purposes. EXPOSURE PROCEDURE
Pigment combinations mere dispersed in three matrices of thv following compositions b y passing twice through a three-roll mill :
W W rosin Methyl abietate Pliolite 9-1
A 87 12 1
Per Cent by Weight B 75 20 5
c 75 6
20
Vchicle A was pigmented at 12% b y volume, B a t Z4YO,and C a i 36%. Plym-ood panels were coated with approximately 10 mils of each composition and exposed at Miami Beach, f l a . , for periods u p t o 26 months. T h e panels were inspected monthly and rated with reference to the percentage surface area remaining free from fouling attachment. When a formulation dropped below SO%, it was no longer considered of interest. I n the tables that follow, the figures indicate the total number of months during whicli