INDUSTRIAL AND ENGIXEERING CHEMISTRY Rassow, B., and Zickman, P., Cellulosechemie, 11, 45 (1930); J. Prac. Chem., 2, 123, 189 (1929). Ritter, G. J., and Fleck, L. C., IND. ENG.CHEW,15, 1055 (1923). Ritter, G. J., and Seborg, R. M., Ibid. To be published. Ross, J. J., and Hill, A. C., P u l p Paper MUQ.Canada, 27, 541 (1929); 29, 569 (1930). Sherrard, E. C., and Beglinger, E., Forest Products Laboratory unpublished report.
Vol. 24, No. 1
(22) Venkateswaren, S.,Quart. J . I n d i a n Chem. Soc., 2, 253 (1925). (23) Von Wacek, Anton, Ber., 63, 2984 (1930). (24) Willstatter, R., and Kalb, L., Ibid., 55, 2640 (1922). (25) Willstatter, R., and Zechmeister, L , Ibid., 46, 2401 (1913).
RECEIVEDOctober 1, 1931. Presented before the Division of Cellulose Chemistry at the S2nd Meeting of the American Chemioal Society, Buffalo, K.T., -4ugust 31 t o September 4,1931.
Seasonal Manganese in a Public Water Supply E. S. HOPKINS ~ N D G. B. MCCALL,Montebello Filters, Bureau of Water Supply, Baltimore, M d .
M
ANGANESE in a
MANGANESE is produced in stored water
hxANGASESE I N SURFACE W A T E R
public water supply containing carbon dioxide by leaching from is generally conceded T h e a p p e a r a n c e of t h e the underlying soil. This carbon dioxide is to be a n u i s a n c e . Low conmanganese trouble only in the produced by the fermentation of organic material centrations produce a grayish autumn a n d its c o n t i n u i n g left on the sides and bottom of a n unstripped black stain on c l o t h e s when until midwinter indicates t h a t l a u n d e r e d and c l o g z e o l i t e the “seasonal turnover’’ of the reservoir. During the autumn “turnozier” this softeners and industrial pressure reservoir water is a contributmanganese is distributed through the water, and, filters, thereby b r i n g i n g comi n g f a c t o r . The c u r v e s in by aeration, is conzierted into an unsfable hyplaints from bottlers, laundries, Figure 2 show this to be true. drated oxide. and dairies. Occasionally the D u r i n g the w a r m s u m m e r The decomposition of dead microorganisms does average consumer protests bemonths the temperature of the bottom w a t e r g r a d u a l l y incause of stains on white enamel not affect this condition materially. Hence fixtures. c r e a s e d from 5” to about 20” copper sulfate treatment is effectioeonly to control As manganese is easily oxiC., the surface water being conthe taste and odor of the water. dized, it usually occurs as an stantly about 3 to 5 degrees The .feasibilily of stripping a reserroir for higher. With t h e b e g i n n i n g unstable hydrated oxide of unmanganese control depends on the cost of the certain composition, p o s s i b l y of colder weather, a b o u t October 1, the surface temperaMna04. It is therefore e a s i l y work as compared with the cost of iron and lime reduced, f r e q u e n t l y giving a ture fell rapidly. The temperacoagulation for a f e w years. tures were soon equalized and f a l s e c o l o r r e a d i n g f o r the a eradual reduction f o 11o w e d , available chlorine determination until final stagnation was reiched, about December 18, ( 5 ) . Unless properly corrected, such a situation can readily cause a dangerous sanitary condition in the control of a water a t 5 ” C. During this period the manganese content of the supply. A recent analytical modification of the test for surface water, which up to this time had remained a t 0.01 manganese in the presence of manganous compounds has p. p. m., increased. This increase persisted until about been published ( 2 ) . Manganese is removed from a public January 1. Using the presence of manganese in the surface water supply in the purification process by using ferrous water as a criterion, the belief that “this overturn continues sulfate and lime as the coagulating medium ( 2 ) . The ferric until a period of fairly stable equilibrium known as winter hydroxide formed a t its isoelectric point, p H 9.2 to 9.4 for stagnation is reached” (3) is correct. Further inspection thir water, will completely adsorb manganous oxideq ( 6 , 7 ) . of the curve shows that the manganese increase occurred only when the temperature fell below 15” C. This proves COKDITIONS AT RA4LTIMORE that the mixing of the top and bottom waters distributed Manganese in excessive amounts is a seasonal phenome- the element through the reservoir. Comparison of the water temperatures from S o . 1 Bridge non in the river water used as the source of supply in Baltimore, Md. The curves in Figure 1, which are weekly aTerages with those from Warren Bridge, 41/2 miles further up stream, of daily analyses a t the plant, show that this trouble occurs having a depth of about 25 feet, indicates that stagnation in the late autumn, reaching a maximum in October of each did not occur a t this point. The results in Table I, Secyear. The investigation here reported was undertaken to tion A, show this in detail. The results in Section 13 confirm discover the cause of this condition and to find a remedy for it. the belief that while manganese is initially found in the mud Weekly samples of surface and bottom waters were col- on the reservoir bottom it is brought to the surface only lected from a point in the 23,000,000,000-gallon reservoir, during periods of “seasonal turnover.” The data for the KO.1 Bridge, approximately 50 feet deep and about 2 miles Paper Mill Bridge Section, with a depth of only about 10 upstream from the intake a t the dam. The surface samples feet, support this opinion. Proof that manganese is obtained from the bottom water were collected in a bottle of 1000 cc. capacity lowered from the bridges and the bottom samples in an apparatus designed is given by the curves in Figure 3. These show clearly that for the purpose (8). KO. 1 Bridge is about 8 miles below the a t certain seasons of the year appreciable quantities of this head maters of the reservoir. Hence the characteristics of element are present in the water. Beginning to increase the water a t this point would be only slightly, if a t all, in- the middle of June, it reaches a maximum by the middle fluenced by fluctuations of contributing stream flow. The of August, remaining fairly high until October. At the manganese conditions a t this spot are comparable to those period of “seasonal turnover,” about October 20, a rapid reduction in the bottom water, with a corresponding increase of a deep pool under climatic and wind effects.
INDUSTRIAL AND ENGINEERING
January, 1932
TABLEI.
CHhRACTERISTICS OF \vATER AT WARRES AND PHOENIX BRIDGES'
BECTION A
I\fONTH
BRIDGE Temperature Deoth Bottom T o o Feet O C . O C , 1930, Aug. 28.5 26 26 Sept. 26.0 25 25 Oct. 24.0 16 17 10 12 Nov. 21.5 5 3 Dec. 19.5 1931. Mar. 24.0 April 26.5 .. , . 18 21 May 28.0 24 23 June 28.5 .Tnlv 2 .~ 7.5 25 29 . a The Warren Bridge location is averages of weekly samples Tq-ARREN
~~
&lILL
g:
ig
This digestion is comparable with that of organic servage material in that it proceeds readily with a minimum amount of oxygen. A discussion (0) of digestion-of-sewage-sludge experiments suggests that the optimum point is a t about 28' C., with a p H value from 6.8 to i.6, and that complete digestion will take place in 30 days. As the usual temperature of the reservoir in summer is between 20" and 30" C., and the p H between 6.0 and 6.8, with the oxygen content very low, it is logical to assume that an analogous digestion occurs on the bottom. This more or less septic state pasily produces excessive amounts of carbon dioxide in conjunction with other resultant gases. il comparison of the characteristics of the water at the Warren Bridge and a t the Paper Mill Bridge locations supports this T ieiv (Table I). S o relation in constituents between the top and bottom waters is observed. If there is an apparent sequence it does not remain for any definite period. The Warren Bridge location presents the effect of contributing stream flow and the other point only that of shallow water. The hydrogen-ion concentrations of the top and bottom waters (Table IV) prove the presence of excessive carbon dioxide. The bottom water was decidedly more acid than the top, oning to adsorption of the gas. It remained so until the period of winter stagnation. Water containing free carbon dioxide will dissolve hydrated mangznous oxides (5, I O ) . Therefore such water
TABLE11. DESSITY OF W.4TER
c.
Gram/cc. 0,997044 0.998203 0,999099 0.999700 0.999965 0.999973 0.999841
90
15 10
5 4
0
The figures in Table I1 explain the suddeii diffusion of manganese through the reservoir. The increase in density when the temperature falls from 25" to 15" C . is much greater (0.2 per cent) than when it falls from 15" to 4' C. (0.09 per cent). With the water quickly cooling from about 2.5' to 15" C. in two weeks, the density changed rapidly in conformity with these ratios, thoroughly mixing surface and bottom waters. This theory is confirmed by the finding that the maximum amount of manganese a t the surface in October is intermediate between that normally found in the two waters. Additional confirmation is given by the dissolved oxygen curve in Figure 3. When the temperature l5
SECTIOX C
BECTION B
BRIDGE hfiNG4NESE AT MANGANESE \ T W ~ R R EBRIDGE N PAPERMILL BRIDGE Temperature WARREN BRIDGEPAPERMILLBRIDGE Dissolved 0 pH Dissolved 0 DeDth Bottom TOD Bottom TOP Bottom Top BottomPH Top Bottom Bottom Top Bottom Feet O C . O C . P. 9 . m. P. 9. m. P. p . m. P. p . m. R Snt. q Sa/ 11.5 0.26 0.01 0.44 0.11 7.0 7.4 30 6.9 7.1 65 8.5 0.57 0.10 0.60 0.32 6.9 6.9 60 6.8 6.9 63 6.5 16 16 1.05 0.24 0.32 0.16 7.1 7.1 68 7.1 7.2 76 7.1 71 . . . . . . 0.43 7.1 0.37 0.33 .. 3.5 . . . . 7.0 79 . . . . . . 0.14 .. 1.20 7.1 2.0 0.14 6.0 .. .. 0.04 .. 0.08 .,. ,.. . . . . . . 9.0 0.02 0.04 . . . . . . . . . . . . 73 74 6.8 7.0 0.05 0126 0.12 7.1 7.4 19 19 0:i3 10.0 67 51 6.8 7.1 6.6 7.5 0.04 0.33 0.09 0.28 11.0 24 26 7.3 62 0.04 0.30 0.13 6.6 7.6 34 6.6 10.0 28 29 0.65 influenced by contributary Ytream flow; the Phoenix Bridge point is comparatively shallow. Figures are monthly PAPER
in the surface samples, is observed. This proves that the water had been equally mixed. That a rapid mixing of these waters would occur is shown by reviewing any table of density in its relation to temperature change. Table I1 gives a few pertinent relationships.
"5
107
CHEMISTRY
r
SI0
1923
1924
1925
1927
'926
FIGURE 1.
PLANT
fell below 15" C. an immediate increase of this element was noted in the hottom water, showing that admixture had taken place. BOTTOM IVATER Except possibly in the colder months, manganese is always present in the bottom water in amounts above 0.1 p. p. m. (Figure 3 ) . This increase occurs after the temperature of the bottom water has been a t 20" C. or above for about 30 days. During this period a progressii-e reduction in the dissolved oxygen is found. This falls to a rniriimum of 15 per cent saturation coincident with the niaximuni temperature of 23" C. The curves in Figure 3 indicate a fermentation of the organic material on the reservoir hottom caused by the action of proteolytic bacteria. The carbon dioxide liberated by this process is suggested as the source of soluble manganese. AIASG.4XESE IS
1928
1979
1930
D4T4, 1923-30 stored over an area of manganous rock deposits, particularly in deep reservoirs, \vi11 under the conditions of fermentation always contain these soluble salts in the bottom water. At the time of "seasonal turnover" this element will be distributed throughout the body and become aerated and the soluble salts nil1 be converted into a hydrated oxide.
CORRECTIOX OF MAXGASESE TROUBLE The yearly curr'e-, in Figure 1 reveal a gradual decrease in the maximum amount of manganese found for each period. I n 1921 a new area of land was flooded. Before flooding the top soil was not stripped or grubbed. This condition produced an abundant supply of organic material for fermentation and subsequent leaching of manganese from the soil on the bottom. The gradual decrease in the peaks of the curve indicate that with the passage of time this fermentation will be checked by silting of the bottom, thereby
INDUSTRIAL AND ENGINEERING CHEMISTRY
108
Vol. 24, No. 1
eliminating the trouble. The absence of silting in 1929 and 1930 by run-off is reflected in the increased peaks for those years. It is believed that with abatement of the drought and with normal rainfall the peaks will show the usual yearly drop. Baylis ( 1 ) originally thought the phenomenon was due to the presence of a particular slime bacteria producing organic acids that dissolved the manganese. I n recent years this organism has not been found in the water thus eliminating i t as the causative agent. The possibility that decomposing microorganisms supply the organic material for digestion was considered. Table I11 presents the relation between periods of maximum organic content and manganese appearance in the water. These figures indicate that the
'PI-
--SURFACE WATER PPPIMn ---SURFACE WATCR "C. BOTTUM WATER'C.
-----
increase in carbon dioxide is not a direct function of microorganism concentration. No indication was found that an extended period of moderate microorganism growth produced excessive fermentable material. It must be concluded that digestion of the excessive organic material from the unstripped reservoir bottom is the underlying cause of the trouble. TABLE111. RELATION OF MICROORGANISMS TO DURATION OF MANGANESE IN WATER YEAR
a
MANGANESE Duration Maximum Duration0 Weeks P. p. m Weeks 52 1.3 12 6 1.4 8 8 0.8 10 8 1.1 8 8 0.8 4 2 1 0.6 12 4 0.7 4 0.9 7
MICROORGANIsMs
Maximum N o . per cc.
Above 0 . 5 p. p. m.
The only permanent remedy for a similar condition would be to remove the growth from the sides and bottom of the reservoir before filling. This is very widely practiced in New England, especially in Massachusetts. Digestion is reduced to a minimum, eliminating soluble manganese as a by-product. Stripping and grubbing a reservoir are expensive. This cost must be compared with that for iron and lime coagulation over a few years. If the cost of such coagulation is prohibitive in plant control for a giyen loca-
POSITIONOF RAYONINDUSTRY IMPROVED. The United States
rayon industry, the most important factor in the world's rayon trade, will start 1932 in a much better statistical position than it entered 1931, according to the Department of Commerce. The slashing of huge stocks that had piled up a t the close of last year is one of the most important reasons for the improvement. Trade figures show that stocks have been cut from approximately 27,000,000 pounds to some 15,000,000,which is practically a normal supply. Following a fairly active demand for rayon prior to October 1, sales have slackened, partly because the demand from weavers
TABLEIV. PH OF WATERAT No. l BRIDGE (Monthly average of weekly samples) DATE PH TOD Bottoma 1930 Mar. 7.2 7.3 April 7.1 7.1 May 7.5 7.0 June 7.5 6.7 July 7.4 6.6 Aug. 7.5 6.6 7.4 Sept. 6.7 7.2 Oct. 6.4 7.2 Nov. 6.0 7.1 Dec. 6.0 1931 May 7.5 6.4 7.6 June -6.0 7.6 July 6.2 7.5 AUE. -6.0 * Bottom samples from 50-foot depth.
LITERATURECITED (1) Baylis, J . Am. Water Works Assocn., 12, 2 2 4 (1924). (2) Enslow, Water Works and Sewage, 79, 184 (1931). (3) Flinn, Weston, and Bogert, "Water Works Handbook," p. 677, McGraw-Hill, 1918. (4) Hodgman and Lange, "Handbook of Chemistry and Physioa," p. 437, Chem. Rubber Co., 1925. (5) Hopkins, IND.EXQ.CHEW,19, 744 (1927). (6) Hopkins, Ibid., 21, 58 (1929). (7) Hopkins, Ibid., 22, 79 (1930). (8) Hopkins, Eng. News-Record, 100, 8 7 0 (1928). (9) Keefer and Kratz, Ibid., 102, 103 (1929). (10) Robinson, Gardner, and Holmes, Science, 50, 4 2 3 (1919).
RECEIVED September 10, 1931. Presented before the Division of Water, Sewage, a n d Sanitation Chemistry a t t h e 82nd meeting of t h e American Chemical Society, Buffalo, N. T.,August 31 t o September 4 , 1931
and knitters has fallen off and partly because of the withdrawal of guarantees. Nevertheless, the trade believes production for the year will slightly surpass that of 1930, which was approximately 115,000,000 pounds. This year's output is estimated a t some 120,000,000pounds, or just short of the 1929 record of 121,566,000 pounds. The domestic textile industry, moreover, is relying less on foreign materials than formerly. Last year imports of rayon yarns were only 5.76 per cent of domestic production, as compared with 12.33 in 1929, 12.47 in 1928, and 19.21 in 1927. The 1931 proportion will be less than 2 per cent.
