Removal of Carbon from Hot Caustic Solution

829 (March 8, 1938). (2) Mark, II., and Wulff, C. (to 1. G. kaiknindustrie A -G.), Ibad. 2,110,833 (March 8, 1938). (3) Mattox, IT. J., and Grosse, .\...
3 downloads 0 Views 350KB Size
INDUSTRIAL AND ENGINEERING CHEMISTRY

832

.4 compvsite yield summary for the series, based chiefly on the distillation method, is shoRn in Table 11, wheie a value of 89.7 weight is given as the ultimate yield of styrene. Deviation from this value did not exceed 3.0 for any one pass by any of the methods. In the fifth pass, where ultimate yield was determined by all three methods, the values xere 91.4, 91.5, and 91.7 weight %, respectively.

r0

ACKNOWLEDGMENT

The authors express their appreciation to Louis S. Kassel at whose suggestion the work x i t h benzene diluent was initiated. LITERATURE CITED

(1) Dreisbach, R. R. (to Dow Chemical Co.), U. S . Patent 2,110,829 (March 8, 1938).

Vol. 38, No. 8

(2) M a r k , II., and Wulff, C. ( t o 1. G . kaiknindustrie A -G.), Ibad. 2,110,833 (March 8, 1938). (3) M a t t o x , IT. J., and Grosse, .\ V.,J . Am. Chem. S O C . , 67, 85 (1945). (4) Mavity, J. AI., Zetterholm, E. E., and Hervert, George, T i m s A m . Inst. C'hem. Engrs., 41,619-28 (1945). ( 5 ) Mitetiel!, J. E., Jr., Am. Inst. Chem. Engrs., Chicago. Dec 16-19, 1945. (6) Smith, 0. H. (to Naugatuck Chemical Co.), U. 9. Patent 1,870. 876 (-4ug. 9, 1932). (7) U. 0. P. Laboratory Test Methods for Petroleurn and Its Products, Method G-84-40, p. G-33. (8) I b i d . , H-44-40, p. H-25. (9) I b i d . , M-14943, p. M-3. (10) Wulff, C., and Roell, E. (to I. G. Farbenindustrie A.-G.), U . S Patent 1,986,241 (Jan. 1, 1935). PREBESTED before the Division of Petroleum Chemistry a t t h o 109th l l e w ing of the ANERICAXCHEXICAL SOCIETI..4tlantic City. pu'. J.

Removal of Carbon from Hot Caustic Solution A. J . LUETTGEN, K. S. WHISLER, i N D H. D. B - ~ ~ ' l \ I . A N P . H . Glatfelter C o m p a n y , Spring Grove, Pa.

Separation of fine, suspended particles from hot, concentrated caustic solution presents a difficult problem. Because of the large volumes of cooking liquor required in a pulp mill, a method similar to the continuous filtration of water is desirable. Investigation revealed that anthracite

r\lTECESSITY

for clarifying hot, concentrated sodium hydroxide solution has led t o the development of a continuous filter system capable of handling a hot caustic solution a t the rate of 100 gallons per minute. The process for recovery of soda ash from the spent liquor from the digesters in operation a t this mill results in the formation of a mixture of sodium carbonate and carbon, known. as black ash. When the sodium carbonate is leached from the black ash, the larger carbon particles settle out and are removed, but the fine particles remain suspended in the sodium carbonate solution which goes to the causticizers. Subsequent to the causticizing operation, the liquor passes through a Dorr clarifier where the calcium carbonate produced by causticizing is removed. Normally this liquor is then ready for use in the digesters. However, m&h of the fine carbon remains in suspension even after passing through the Dorr clarifier. So much carbon is present that the liquor is black and

coal as a fitter medium auccessfully removes finely divided particles and is sufficiently inert to hot, concentrated sodiuni hydroxide. Filters in operation are remo\ing ' finely di\ided carbon (largely under 15 microns) from 11 % caustic solution at the rate of 100 gallons per minute.

opaque a3. Figure 1 shows. Unlt!SF it is removed before going to the cooking process in the digester, carbon is introduced into the pulp where it lowers the brightness of the pulp and eventually of the finished paper. This, then, presents the problem of removing finely divided, suspended carbon from a hot sodium hydroxide solution in a manner capable of handling continuously a flow from the clarifier of about 100 gallons per minute. The amount of carbon in suspensiori varies from day to day, probably e w n from hour to hour; it seldom exceeds 8 pounds per 1000 g a l l o n s , rarely falls below 1.5 p o u n d s , a n d a v e r a g e s about 3 pounds. The t'emperature of the hot caustic in which the carbon i p tiiFpcraet1 is about, 160" F., and its concentration is 113 grams per liter of total alkali expressed a: SnOH. This is about, a 2.8 molar solution of sodium hydroxide or an 11% solution. Tlie size of the carbon particlo? \vhiich remain suspended in the liquor varies greatly. More t h m

INDUSTRIAL AND ENGINEERING CHEMISTRY

August, 1946

I

I

I

77me Figure 2.

I

2.0

I

I

30

/;.l/f/OU.J

Effluent Rate and Efficient? with 0.43-1Im. inthrafilt

50% of the dispersed particles are under 15 microns, and about 90% are under 30 microns. Several methods were considered for removing this carbon: (a) Simple settling was shown to be inadequate. ( b ) Settling aided by flocculation was found to be effective but economically impractical; further investigation on flocculation and flocculating agents is contemplated. ( c ) Filtering was found experimentally to be feasible and has been put into successful plant operation. Because most of the commercial filtering media are rapidly attacked by hot caustic solutions of the concentration encountered here, the first problem in filtering m-as to select a suitable medium. It seemed probable that Anthrafilt would be sufficiently inert t o the caustic liquor, Anthrafilt is Pennsylvania anthracite coal properly selected and ground to serve as a filtering medium

bed so that there ra,s a continiious, slow di:scharge from the overflow at the top of the pipc. AAftereach run the filter was backn-ashed with water a t the maximum vclociity possible without loss of coal. After preliminary tests with the experimental filter, it was decided to keep a constant setting of the effluent valves and periodically measure effluent rate and filter efficiency until either dropped to an undesirably low point. Space available for mill installation indicated that filtering rate should not drop below 1 gallon per square foot per minute. Because of the extreme fineness of the suspenlled particles to be removed, it was thought that a fine filter medium would be required, Hence tests were made on two Anthrafilt samples (effective size 0.43 and 0.59 mm.), both of which were smaller than ordinary commercid grades. A s Figure 2 shows, with effective size 0.43 mm. both the effluent rate and efficiency drop off rapidly. Similar results were obtained with effective size 0.59 mm. This was taken t o indicate that, rather than penetrating through the filter, the fine carbon particles were accumulating in the upper regions of the,bed and quickly clogging the filter. Visual examination of the filter bed revealed unmistakable evidences of channeling. This would explain the appearance of carbon in the effluent early in the filter run. Filter medium of 0.70-mm. effective size was tried with marked improvement (Figure 3). I n both these experimental runs the filter bed was 36 inches deep. From the experinlental results it was predicted that with an effective size of 0.70 mm., filter runs in excess of 5 hours would be possible before backwashing would be necessary, with effluent essentially free of macroscopically visible suspended particles. The filters subsequently constructed and nom in plant operation are giving satisfactory runs of 12 hours before backwashing. These filters need virtually no supervision, which leaves the operator free for other duties except during the short backwashing period. Other than replacement of small amounts of Anthrafilt lost during backwashing, the plant installation has required no repairs in more than ten months of operat,ion. APPLICATIONS

Because Anthrafilt is essentially carbon, it is inert chemically to acid, alkaline, and oxidizing solutions. In addition to its ordinary use in water and sewage filtration, its properties are suited to the construction of long-life continuous filters to remove very fine suspended material from corrosive solutions.

U,@. Turner and Scott (3) report that, after the initial loss amounting t o about 5% of the total weight of anthracite, 1% sodium hydroxide has no further action on the coal. Before the ability of Anthrafilt to remove the fine carbon particles was tested, measurements were run to give an indication of the solvent action of 10 and 30y0 solutions of sodium hydroxide on anthracite a t 160" F. , The behavior of 10 and 30Y0 solutions was nearly identical, and after 48 hours of exposure to the solvent action of the caustic, the anthracite coal had lost less than 257, of its initial weight. Further decrease in weight was negligible. This shows markedly less solvent action than Turnel, and Scott obtained with lyO sodium hydroxide, but their measurements were made at elevated pressures and temperatures of 270" F. and higher. An experimental filter was constructed from a cast iron pipe, 4 feet long and 4 inches in diameter. This was fitted a t the bottom with an interior perforated plate and an cxt,erior valve to control the effluent. The filter bed was constructed along the same plan tw a sand filter for water purification. 911 experimental runs viere carried out under constant head, maintained by keeping sufficient liquor above the

L8

'$

a33

-

-

A6-

$ x4 ? a: $ 4

t f

LO

a8

-

-

LO

2.0 30 40 Z h e /A HOOPS

40

20 30 40 77ine /.+Him

XU

Lzzh

s /oo .E .'t 80 6< 60

Figure 3.

40

Effluent Rate and Efficiency w i t h O.iO-3Im. Anthrafilt

INDUSTRIAL AND ENGINEERING CHEMISTRY

a34

Possible applications are (1) filtration of concentrated sodium hydroxide or sodium carbonate solutions in the paper, rayon, and textile industries; (2) filtration of bleach liquors; (3) removal of suspended material from strong acids, as in the refining of titanium dioxide; (4)any plant operation where it is necessary to filter caustic and corrosive solutions and where a long-life continuous system requiring a minimum of attention is desired.

Vol. 38, No. 8

ACKNOWLEDGMENT

The authors express their appreciation for the helpful suggacltiom and criticisms given them by the late H. G. Turner and by ** witt Hutchison. LITERATURE CITED (1) Turner, H.G., Isu. Exa. CHEM.,35, 145 (1943). (2) Turner, H, G , , J , Am. water Assoc., 36,431 (1944) (3) Turner, H.G., and Scott, G. S , Combustion, 5 , 23 (1934).

Acetic Acid-Ethyl EtherWater Svstem J

MUTUAL SOLUBILITY AND TIE LINE DATA COLEMAN J. MAJOR1

AND

OSCAR .J. SWENSOX,

>Iutual solubility and tie line data are presented for the system acetic acid-ethyl ether-water. Data at 25' C. were determined using reagent and commercial grade ether. 4dditional data were obtained for commercial ether at 20 C. The tie line data have been successfully correlated by each of four different methods. Data on the water side of the mutual solubility curve were obtained by a new method based on the difference in foaming tendency between a homogeneour and a heterogeneous solution of the three components.

I

9 T H E course of a laboratory investigation on liquid-liquid

extraction, mutual solubility and tie line data xere required on the system acetic acid-ethyl ether-water. Although a number of investigators have studied this system (1, 3, 6-15, 16, 18, 19, ?O), the published data are not sufficiently extensive t o permit mnstruction of a complete ternary diagram. The present study &vasundertaken for the purpose of obtaining the necessary data with both reagent and commercial grades ethyl ether. The equilibrium data for the system are presented in Tables I and I1 arid the ternary diagrams are given in Figures 1 and 2. The mutual solubility curve of Figure 1 for commercial ether lies below that of the reagent grade because the commercial ether used in this investigation contained, as an impurity, approximately 3.5 weight 70 ethyl alcohol. Figure 2 shows that very litI!(! difference exists between the data for 25" C. and for 20" C. 1

Cornell University, Ithaca, N . \ .

The tie line data of Table I1 indicate that, a t low concentretiona of acetic acid, the acid concentration in the ether phase is Ion-er than that in the water phase; a t high concentrations of acetic acid the reverse is true. This is illustrated graphically by Figures 1 and 2, where the tie lines actually undergo a rcvrssl io dope as the acid concentration is increasrd. GRADE OF MATERIALS

ACETIC ACID. Baker and hdamson reagent, glaviiil acetic acid was used. Direct titration with standard alkali and phenolphthalein indicator showed the acidity to be 99.8% hy weight. Density data on the acid a t various temperatures are prwented in Table 111. At 23' C. the demity of the arid wn.5 1.0439 grams per ml. n-hereas the accepted literature value is 1.0440 for tJhe pure acid. ETHYLETHER.The reagent grade was obtaitir~lfrorii onepound bottles labeled "Ether Anhydroris LIwck". Th cations indicated that principal inipuritit:s ~verc 0.1 alcohol and 0.05yc n-ater. The comnitwial grade wa: from 5-gallon metal containers l a h c l ~ d"Ether [*.S.l'. LIvrck". It was an alcohol derivative containing itpproxinia t vIy :3.5 weight % ethyl alcohol. Density data on the two grades are prevcntcd in 'l'abltr 111. The density of the reagent grade at 23' C. ivas 0.7071 gram per nil. as compared n*ith the accepted literature valuc: of 0.7078 for the pure ether. 'ATER. Ordinary distilled water from the laboratory supply

Present address, Sharples Chemicals, Inc.. Kyandotte. Mich

A C E T I C ACID ACETIC ACID

-Reagent ---

2 0 0 c. 25' C.

Ether

Commercial Ether

" WATER

Y

"

"

"

"

WATER ETHER

Figure 1. Mutual Solubility Curves and Tie Lines for Acetic Acid-Ethyl Ether-Water System at 25" C.

ETHER

Figure 2. Mutual Solubility Curve and Tie Lines for Acetic Acid-Commercial Ether-Water System at 20" and 25" C.