Lime Treatment of Waste Pickle Liquor - ACS Publications - American

particularly spent pickling liquors, with the object of demon- which would require substantial quantities will be forced to await strating the means w...
0 downloads 0 Views 588KB Size
INDUSTRIAL AND ENGINEERING CHEMISTRY

February 1947

After the products are collected in storage tanks, the milling and packaging steps that follow aTe substantially the same as those used in conjunction with other methods of fatty acid manufacture. Since the Emersol process involves the use of a flammable solvent, cert,ain safety factors must be considered. The narrow explosive range of methanol, however, coupled with the fact that all of the operating units are enclosed, precludes the possibility of a combustible level being reached in the atmosphere. In addition, safety valves and alarms are dispersed throughout the plant to nullify any excess pressure rises and indicate any operational irregularities. The plants are designed for 24-hour-a-day operation, except for scheduled shutdowns for routine maintenance. Comparison of the operating costs (Table 111) with those for mechanical pressing indicates that Emersol's are about 65% less (8). This difference is due mainly to the higher labor figure for the batch method as well as its additional expense of liquid acid recycling and purification, operations not required in the continuous method of separation.

131

ACKNOWLEDGMENT

The author wishes to acknowledge the helpful cooperation of V. J. hfuckerheide and L. D. 91yers of Emery Industries in the preparation of this art,icle. LITERATURE CITED

(1) Bailey, A. E.; "Industrial Oils and Fat Products", pp. 434-6, New York, Intersoience Publishers, 1945.

iii :!:::; ::: ~~~~~: ~ ~ B ~a.c.,~U,s,patent

2,362,160 , (1944). ( 5 ) Hilditch, T. P., "Industrial Chemistry of Fats and Waxea". 2nd ed., p. 7, London, BailliBre, Tindall and Cox, 1941. (6) Ibid., pp. 275-6. (7) Kistler, R. E., Muckerheide, V. J., and Myers, L. D., Oil & Soap, 23,146-9 (1946). (8) Ihid., 23,150 (1946). (9) Myers, L. D., and Muckerheide, V. J., U. S. Patent 2,293,876 (4)

(1942).

Ibid,, 2,298,501 (1942). (11) U. s. Dept. of Commerce, Repts. of Industry Div., Fata and Oils Unit, Bur. of Census, 1945.

Lime Treatment of Waste Pickle Liquor Richard D. Hoak, Cliflord J . Lewis', Charles J . Sindlinger, and Bernice K l e i n MELLON INSTITUTE, P I T T S B U ~ G H , PA.

T

HE first, paper in this

The current scarcity of high-calcium lime suggested Dolomitie limestone occurs series described rapid the studies reported in this paper, which describes methabundantly and is calcined analytical methods for deods for attaining equivalent results with the less reactive, to lime of chemical quality termining the available alkabut more readily available, dolomitic lime where this near all major centers of linity of limes and limestones material must be used for waste pickle liquor treatment. steel production. In most and the total sulfate ion cases, especially where its in waste pickle liquor (1). higher basicity is the princiThe second article discussed certain factors having economic pal factor, dolomitic lime can be delivered t o consumers importance where high-calcium limes and limestones are used to a t lower cost than similar high-calcium products. A contreat such waste liquors ( 2 ) . sideration of these factors emphasized the desirability of Dolomitic limes were not studied in detail in the previous work. studying the reaction between dolomitic limes and acidic wastes, The demand for high-calcium lime is now so great that new uses particularly spent pickling liquors, with the object of demonwhich would require substantial quantities will be forced to await strating the means whereby this low-cost basic agent can be employed most effectively. augmented production. This situation results in part from the fact that high-calcium limes have long enjoyed the select chemical EXPERIMENTAL WORK , and metallurgical markets xhich, expanded during the war, continue to tax the productive capacity of the industry. On the Kine types of lime, two high-calcium and seven dolomitic, supother hand, circumstances during the war years curtailed cerplied by an Eastern Pennsylvania manufacturer, were investitain choice markets for dolomitic lime; this condition, coupled gated. The samples were characterized as follows: with competitive factors, has resulted in a moderate surplus of dolomitic lime a t present. It appeared reasonable t o assume 3. High-calcium lime 4. Shaft-kiln dolomitic lime that, if the over-all expansion and recovery of lime markets in 6. Rotary-kiln dolomitic lime general permit any surplus of lime to accumulate in the next 6. 60-40 blend of shaft- and rotary-kiln dolomitic lime several years, such a surplus will be of the dolomitic material. 7. High-calcium lime hydrated a t atmospheric pressure Stream Dollution increased during the war through the es8. Dolomitic lime hydrated a t atmospheric pressure 9. Dolomitic lime hydrated a t 30 lb./sq. in. tablishmeni of new industries, enlarged production by old ones, 10. Dolomitio lime 'ydrated at 30 'b /Sq in and air-floated lack of materials for constructing adequate treatment works, 11. Dolomitic lime hydrated a t atmospherlc pressure and air-floated and depletion of the engineering staffs of public health agencies. In all industrial sections of the country health authorities have and 11 Were used in most of the samples 4,j, now embarked On aggressive campaigns to reduce stream polluwork because they represent the usual types generally available. tion by requiring satisfactory treatment of municipal and indusThe samples were analyzed, and then basicity factors were calcutrial wastes. Where treatment of acidic wastes is involved, lime lated from the analyses and determined experimentally normally provides by far the cheapest source of basic agent. sulfuric and hydrochloric acids. The resulting data are given in 1

Present address, Warner Company, Philadelphia, Pa.

Table I.

INDUSTRIAL A N D ENGINEERING CHEMISTRY

132

VOl. 39, No. 2 REACTION K 4 T E

TABLEI. As.ii.usi.:s A

.

1

analysis,

AXD

c 1 . ,

v0

SiOn RzOa CaO b1gO lloisture Total

+

3 1.51 0.88 94.17 0.65 COz 2 . j 8 99.79

4 2.69 0.70 55.03 40.70 0.68 99.99

5 3.15 0.70 54.95 40.10 1.05 99.95

BASICITYFACTORS OF hT1mLIMES Sample Sumber 7 8 4.13 1.12 2.72 0.53 0.96 0.48 53.99 71.40 45.04 40.52 0.59 33.02 0.96 p5.90 18.93 100.13 99.96 100.19 6

Basicity factor" Calcd.

0,9508 1.120'2 1.1109 1.1071 0.9515 1.0837 1.0855 1.0694 HC1 0.9817 1.1189 1.1351 1,1027 0 Basicity factor = grain equivalent CaOIgram sample. HCl by boiling 15 minutes: others, 30 minutes.

Hz80r

0.7222 0.7045 0,7225 factors and

7

9 2.18 0.57 41.77 29.73 zj.15 99.40

10

11

1.02 0.60 42.01 29.69 99.86

45.90 32.76 99.89

26.54

1.55 LO2

18.67

Where waste pickle liquor is to be treated n-ith linie, the basicity fact,orR i a guide to the quantity required, hut reaction rate is a factor ivhich may be of first importance in some cases. Dolomitic lime i-; avail-

0.912ti 0.8339 0.8357 0,9275 0.8611 0.8302 0.5410 0.8876 0.8952 0.8364 0,8389 0.9115 HzSOd factor for sample 3 determined

It is desirable to point out that, although the experiinerital data to be reported apply equally well t o all dolomitic limes, Eastern Pennsylvania dolomites are eomen hat lees pure than those from other sources. A dolomitic quicklime from Ohio, for example, may contain approximately 0.8% silica and 0.2% R2O3; in consequence, this lime Lyould contain a slightly higher proportion of calcium and magnesium oxides. This difference will be reflected in the corresponding hydrate, depending upon the extent to which air separation removes the impurities. Determination of basicity by calculation from chemical analysis may be misleading because analyses do not reveal the manner in which the silica and RaOs are combined in the product. The availability of the calcium and magnesium oxides is reduced in proportion to the degree t o ahich these impurities have fluxed during calcination. Other minor factors, such as slaking quality and distribution of the impurities between the calcium oxide and magnesium oxide components, cause uncertainties m-here basicity factors are calculated. The rate of development of basicity in sulfuric acid is shown graphically in Figure 1. A basicity factor determined by boiling a sample in an excess of hydrochloric acid for 15 minutes may be regarded as an ultimately available basicity; these values for each sample are shown on the right-hand margin of the figure. For practical purposes the full neutralizing value of these limes is available in about 15 minutes n-ith sulfuric acid under the conditions given; but where sufficient time is allowed, it may approach the hydrochloric acid ultimate. These curves are of particular significance \There the relative delivered cost of the limes is considered.

as quicklime, atnlosphcr.ie hydrate, and pressure hydrate. The magnesia in pressrirehydrated dolomitic limes is almost completely converted to the hydroxide, in contrast ivith atmospheric hydrates where little, if any, of the magnesia is hydrated; this product has been developed within the past six to eight years. Each of these types of lime react,s v-it,li pickle liquor at a different rate. If t,he full basicity of quicklime is to he realized, it must be properly slaked in Tvater before use; likewise, the hydrates must be thoroughly wetted and added as an aqueous slurry. The rate of precipitation of ferrous iron is a function of the rate at which hydroxyl ions can be supplied a t a pH of about 8.5. hi illustration of the rate a t vihich dolomitic limes react vith an acid solution, compared lyith high-calcium lime, a 0.1. .A- solution of sulfuric acid was treated wit,h a 2% excess of several limes a t room t,emperature. The results (Figure 2 ) shoi\- clearly the differences in reaction rates. Similarly, a dilute pickle liquor (10.!% grams Fe++and 38.68 grams S04--- per liter) \vas treated with slurries of the same agents (Figure 3). The rate of removal of iron from a pickle liquor by these agents was traced by treating a pickle liquor (56 grams Fe+', 180 grams Soh-- per liter) a t room temperature with a 5% excess of the limes slurries. The suspensions were stirred const'antly (except where allowed to stand overnight) with electric mixers,

IO

9

8

Z? x 3

c

1/ .0. I t

I 6 IL

0 I p 5

4

3

2

0

I

I

20

40

I

I

60

60

I

100

I

I20

I

I

ULT

I

I

5

10

I

I

15 20 TIME, MINUTES

1

25

TIME, MINUTES

Figure 2.

Figure 1. Basicity Factors of Limes

1

30

Reaction of Limes with 0.1 A' Sulfuric Acid at 25' C.

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

.February 1947

133

TABLE11. RELATIVE RATE O F REACTION BETIYEEX CALCIVM AND MAGKESIUM OXIDES BND HYDROCHLORIC ACID Excess Lime, % '

9-

Grams per Liter hIg

Ca + -

+ +

hiolal Ratio, Ca/hlg

pH of Filtrate

8PH

7-

6A

0

'4

0

'4

D

8'

4 5 IO

"

II

v I

I

I

I

I

I

I

TIME,MINUTES Figure 3.

Reaction of Limes with Pickle Liquor a t 25" C. (pH ws. Time)

samples were withdrawn a t intervals and filtered, and t h e filtrates analyzed for iron. The reaction rate was negligible when the suspensions stood quiescent for 16 hours The data are shomn graphically in Figure 4. RELATIVE REACTIVITY OF CaO AND MgO

difference in reaction rate does exist. In this light it may properly be assumed that this condition is even more significant where sulfates are involved because the removal of calcium ions, by precipitation as a sulfate, increases the driving force for solution of calcium oxide, whereas the reverse is true for magnesia. In addition to the driving force for the solution of magnesia decreasing as the reaction proceeds t o completion, the magnesia which remains a t any time is less reactive than that which reacted previously. The explanat,ion is that it is difficult to calcine dolomite without overburning the magnesia. The double carbonate of true dolomite begins to decompose a t about 725" C. into calcite and magnesite ( 3 ) . This point is higher than the decomposition temperature of magnesite (620 C.) , TThich immediately decomposes into magnesium oxide and carbon dioxide. Calcite does not decompose below about 900' C.; owing to the intimacy of contact between the magnesia and the lime, attainment of complete decomposition of calcium carbonate usually results in moderate to severe overcalcination of the magnesia and thereby diminishes its reactivity. These considerations make clear the low rate a t which dolomitic limes react, as compared with the high calcium material, and suggest the means whereby the reaction rate can be increased. Several practices, and combinations of them, are proposed to permit the t,reatment plant operator t o take advantage of the lower neutralization cost of dolomitic limes and their greater availability-namely, use of excess lime, increasing the rate of oxidation of the ferrous hydrate, and raising the temperature. USE OF EXCESSLIYB. The decrease in reactivity of the niagnesia component of dolomitic lime as less and less of it remains brings out the fact that the reaction rate could be increased by employing an excess .of the alkaline agent'. A moderate excess can be used a t the same cost as a stoichiometric quantity of highcalcium lime because of the lower cost of dolomitic lime per unit of basicity.

As a partial explanation for these differences in reactivity it was assumed that the calcium oxide component of dolomitic lime reacted much more rapidly than its magnesium oxide fraction, and that the least soluble and least basic portion of the lime would, in consequence, be required t o precipitate iron a t a pH value where the solubility of magnesium oxide is very low indeed. I n other words, since the reaction is ionic, the driving force for the production of hydroxyl ions becomes almost vanishingly small as the reaction approaches completion. The relative rate of reaction of calcium and magnesium oxides was determined as follows: A solution of hydrochloric acid in a beaker partly submerged in iced water was treated with an excess of dolomitic lime slurry. Just a t the point where the pH of the mixture began to rise, a sample was withdrawn by inserting a 2-cm. tube closed a t one end with a piece of filter paper between two layers of filter cloth and applying vacuum to the open end. iinalysis of the clear filtrates yielded the results presented in Table 11. These findings indicate the correctness of the assumption that the calcium oxide component of dolomitic lime reacts more rapidly than its magnesia. It would be expected that the ratio of calcium to magnesium would increase more or less uniformly with increasing excesses of lime. The accurate determination of the ratio, however, is a matter of some difficulty because a measurable time is required to withdraw a sample, and the rate of reaction in hydrochloric acid, even a t 0 O C., is so rapid that to obtain a sample while the difference in reactivity persists is indeed arduous. For this reason these data must be regarded merely as a Figure 4. Reaction of Limes with Pickle Liquor at 25' C. (Iron Removal us. Time) qualitative indication that a

134

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

Vol. 39. No. 2

* - O r EXCESS

TIM E, HOURS

Figure 5. Effect of Excess Lime i n Pickle Liquor Treatment a t 25 O C. Figure 7. Effect of Increased Temperature, Excess Lime, and Rapid Agitation in Treatment of Pickle Liquor

A

N0.4

0

No.5

No. IO 0 No I I

agitation. This figure combines the effect of the three variables. (excess lime, agitation, temperature) lvhich are important in t h r efficient use of dolomitic lime for t,reating pickle liquor. Corn-. parison of Figures 7 and 4 illustrates the improvement in rcaction rate which can be attained where proper attention is given. these variables. SLUDGE SETTLING RATES

Where there is a limited area for lagooning the slurry from lime. treatment of pickle liquor, the sludge settling rate is quite imTIME, HOURS portant. The fact' that a large fraction of dolomitic lime does not forin an insoluble sludge in pickle liquor treat,ment has frequent,ly Figure 6 . Effect of .4gitation Rate in Pickle Liquor been cited as an advantage for tmhematerial. Figure 8 shone the Treatment a t 25' C. settling rates of the sludge from the samples v-hose reaction rate!: were presented in Figure 4. These curves demonstrate the wide variation in the settling average pickle liquor treated .il-ith 0, j,alld 207, exrates of t'he sludges produced bp the type; of lime studied. In cesses of several dolomitic limes a t roo111 temperature jvith modgeneral, sludges produced hy quicklimes settle fast'er than those crate agitation. The results \yere plotted in terms of iron prestirring time. A trpical set of curves is shoTTn from hydrates, and high Calcium sludges faster than dolomit,ic. cipitation It was shown previously ( 1 ) that t'he settling rate of sludges will in Figure 5. Similar families of curves ]%-ereobtained for other be approximately the same as the settling rate of the lime sussamples which indicate roughly that, for every 5% excess lime pended in Tvater. There is no explanation for the curious hecompletion of the reaction occurs about an hour soonel'. havior of the sludge from sample 6, which settled considerably ISCRE~~SING RATEOF OXID.~TION. Any means JThich \Till tend faster t'han either of the limes from vhich it LTas compounded. to increase the hydrogen ion concentrat,ion in the slurry It is unfortunate that increasing the reaction rate by increas-. increase the driving force for solution of magnesia. Where fering the oxidation rate reduces the settling rate of the sludges rous hydrate oxidizes, the pH of t,he mixt,ure falls and the rate of formed. This effect is shown in Figure 9. It is true that dolooxidation can be increased quite simply by mixing the slurry mitic lime sludges contain a laxer total weight of dry solids than more rapidly to entrain more air. their high-calcium counterparts. The dry solids in the sludges and rapid agit,ation TTas cornpared by treatThe effect of of Figure 8 are as f o h ? - s : ing an average pickle liquor (56 grams F e + - and 178 grams Sod-- per liter) with a 5% excess of several dolomitic limes. .4 Dry Solids, Dry Solids, given slurry was first mixed.by a laboratory stirrer running at Sample G . / L . Original Slurry Sample G , l I , , Original Slurry low speed, and the experiment was repeated at a high rate of 120.0 8 88.5 agitat!ion. The results (Figure 6) illustrate clearly the marked 4 91 1 9 87.5 89.5 10 99.0 5 effect of mixing on reaction rate. This effect is part,icularlp sig7 117.7 11 107.4 nificant in the case of the pressure hydrate; this type of lime is generally less reactive than the other varieties investigated. Where slurries lnust be lagooned, holvever, it is the final bulk Obviously, provision for the more efficient aeration of the slurry sludge mhich is important, llot its dry solids content, with diffused air would increase the reaction rate still further. EFFECTOF TEMPERATURE. It was to be expected that the COST O F LIME PER BASICITY UNIT reaction rate would be increased by increasing the temperature. The lower cost of dolomitic limes per unit of basicity, as coniAs steel is normally pickled in sulfuric acid at 80 a t o 105" C., it, is pared with high-calcium, has been mentioned several times. It usually feasible t o treat the hot liquor. The effect of temperature is desirable to examine this statement in terms of actual cost is illustrated in Figure 7 , which presents the data from treatment of an average liquor with a 5% excess of several limes a t rapid dollars. Pressure hydrates are not included in the analysis be-

TABLE111.

COSTOF AVAILABLE CALCICM OXIDEEQUIVALENT

100

PER T O N OF LIME

Delivery Point

135

INDUSTRIAL A N D E N G I N E E R I N G C H E M I S T R Y

February 1947

High-Ca

Equiv. CaO

Doloinitic

Equiv. CaO

Philadelphia Pittsburgh Cleveland

Quicklimes, Carloads in Bulk $9,55 $9.81 $10.31 9.95 10.42 9.92 9.45 10,40 9.90

$8.80 9.17 8.71

Philadelphia Pittsburgh Cleveland

Hydrated Limes, Carloads in 50-Lb. Bags $10.05 $11.31 $16.04 16.20 11.45 11.42 16,52 10.95 11.65

$11.71 13.09 12.51

90 bp W

5

2 80

3

cause this study has revealed them to have the lowest basicity and reaction rate of the limes investigated; in addition, they cost more per ton. Pressure hydrates were developed with the specific object of improving physical rather than chemical properties of the hydrate-e.g., plasticity. Table I11 gives typical quotations, as of February 1946, for four limes, delivered in carlots to the destinations indicated. Such quotations include both the base price of the lime and t.he freight charges. The prices have been divided by the corresponding sulfuric acid basicity factors of the materials to show the cost of available calcium oxide equivalent per ton. Table I11 makes it plain that basicity purchased as dolomitic lime is cheaper than high-calcium lime basicity by an average of about 14% for quicklimes and 24y0 for hydrates. The cost advantage of dolomitic lime should, in most cases, overcome. its disadvantage of lower reaction rate and place it on a more or less equal footing with high-calcium lime, through the use of excess agent and more rapid agitation for a slightly longer reaction time. SUMMARY

The increasingly aggressive attitude of state and federal agencies in demanding reduction of stream pollution through waste treatment, the cheapness of lime as a neutralizing agent for acidic wastes, and the continuing shortage of high-calcium lime led t o an investigation of means whereby dolomitic lime could be substituted economically for its more reactive counterpart. The lower rate of reaction of dolomitic limes as compared with high-calcium limes under the same conditions is illustrated graphically. Two related factors are largely responsible for the lower reactivity of dolomitic limes: low solubility of magnesium oxide and decreased reactivity of this component through overcalcination. Experimental data indicate that the calcium oxide reacts a t a substantially greater rate than the magnesium oxide in dolomitic lime; in the treatment of an acid solution of ferrous sulfate this necessitates completing the reaction with the least reactive portion of the lime a t a pH where the solubility of magnesium

> W

R:i?APID, S=SLOW MIXING 50

4

0

I

I

I

I

20

40

60

80

TIME, HOURS

Figure 9.

Effect of Mixing Rate on Sludge Settleability

oxide is very low indeed. In addition, the magnesium oxide remaining a t any time is less reactive than that previously consumed by the reaction. The reaction rate of dolomitic lime can be increased by employing an excess of the agent. Very roughly, and within limits, each 5% excess reduces the reaction time about an hour. Raising the temperature of the mixture increases the reaction rate considerably. Advantage can be taken of this factor in practice by treating pickle liquor before it has an opportunity t o cool from pickling temperature. Increasing the rate of oxidation of the ferrous hydrate in the mixture by stirring it more rapidly, or by the introduction of diffused air, markedly increases the reaction rate. Dolomitic lime treatment of waste pickle liquor results in a sludge which usually is bulkier than that from high-calcium lime treatment even though it has a lower content of dry solids. This consequence may be a disadvantage where only a limited area is available for lagooning. The basicity advantage which dolomitic lime holds over highcalcium, however, places the two agents on a relatively equivalent basis for pickle liquor treatment. Where considerable volumes of liquor must be treated, the saving in cost of alkaline agent will permit the use of several expedients t o increase the reaction rate without exceeding the cost of the more reactive high-calcium lime, ACKNOW LEDGMEhT

This work was undertaken as part of the research program of the Stream Pollution Committee of the American Iron and Steel Institute through the Multiple Industrial Fellowship it has sustained a t Mellon Institute since 1938. LITERATURE CITED

(1) Hoak, R. D., Lewis, C. J., and Hodge, W. W., IND.ENG. CHEM.,36, 274 (1944).. (2) Zbid., 37,553 (1945). (3) Sahallis, A., U. S. Bur. Mines, Circ 7247 (1943). PRESEKTEDbefore SETTLING TIME, HOURS

Figure 8.

Lime Sludge Settling Rates

the Division of Wate Sewage, and Sanitation Chemistry a t the 109th Meeting of the AMERICANCHEMICAL SOCIETY Atlantic City, N . J .