The Inhibiting Action of Minute Amounts of Sodium

Jun 27, 2018 - a few minutes. It was found that the number of particles observable under these conditions increased to a maximum in about 5 min. and t...
1 downloads 0 Views 2MB Size
652

T. f. BUEHRER AND R. E'. REITEbfEfER

T H E INHIBITING ACTION OF MINUTE' AMOUNTS OF SODIUM HEXAMETAPHOSPHATE ON T H E PRECIPITATION OF CALCIUM CARBONATE FROM AMMONIACAL SOLUTIONS. 111 THE PROCESS, WITH SPECIALREFERENCE TO FORMATION OF CALCIUM CARBONATE CRYSTALS

MECHANISM OF

T. F. BUEHRER

AND

THE

R. F. REITEMEIER

Department of Agricultural Chemistry and Soils, University of Arizona, Tucson, Arizonu Received August 94, 1030

In a previous paper by the authors (14) it was shown that the inhibiting effect of metaphosphates on the precipitation of calcium carbonate manifests itself as either a retardation of precipitation or its complete prevention, depending upon the properties of the reacting components present in the solution, primarily those of calcium bicarbonate and sodium hexametaphosphate. Chemical studies designed to explain the mechanism of this process have shown definitely that it is not due to some mysterious interaction between metaphosphate and bicarbonate ion, or carbonate ion, or the ammonia. Nor does it affect the possible interactions between these constituents. The minuteness of the concentration of metaphosphate which is effective in this process would appear to eliminate complex formation between calcium ion and the hexametaphosphate as a possible explanation, since only an extremely small amount of it is necessary to prevent precipitation completely. The quantitative evidence obtained in these studies has proved, moreover, that the primary precipitation reaction, represented by the simple equation Ca*

+ HCOB + OH- = CaCOds) + &O

is not affected by traces of hexametaphosphate. The only other plausible basis for explaining the mechanism must be sought either in the effect of hexametaphosphate on the nature of the solid phase which separates out, or in the conditions attending the crystallization process. Preliminary experiments had shown that, when calcium carbonate was precipitated from calcium bicarbonate by the addition of ammonia, the crystals formed rapidly, did not adhere to the walls of the vessel, and settled readily. If, however, the same precipitation was carried out in the presence of one part per million or less of hexameta1 Presented before the Division of Physical and Inorganic Chemistry at the Ninety-sixth Meeting of the American Chemical Society, held at Milwaukee, Wisconsin, September, 1938. See footnote to Paper I of this series for additional information.

PRECIPITATION OF CALCIUM CARBONATE.

I1

553

phosphate, that is, below the level necessary completely to prevent precipitation, there was no immediate precipitation but, after standing several hours, clear, brittle crystals were formed which adhered firmly to the sides and bottom of the precipitating vessel. Microscopically these crystals were strikingly different from the well-known calcite rhombs. It was concluded that the metaphosphate must in some manner interfere with the crystallization process. The problem resolves itself into two questions: ( I ) Does the metaphosphate stabilize calcium carbonate particles of colloidal dimensions and prevent them from aggregating into macroscopic particles, or ( 2 ) does it interfere with the subsequent growth of the crystal from nuclei that were formed in the normal manner? The present paper sets forth a study of these two questions. ULTRAMICROSCOPIC STUDIES

Preliminary experiments indicated that calcium carbonate suspensions formed in the presence of metaphosphate did not show the Tyndall effect. More conclusive quantitative evidence was secured with the slit ultramicroscope, on mixtures containing 50 p.p.m. of calcium, 155 p.p.m. of bicarbonate ion, 560 p.p.m. of ammonia, and sodium hexametaphosphate in concentrations varying from 0 to 5 p.p.m. The preparation of the compounds used in making up these solutions has been described in the preceding paper (14). These solutions had been prepared from water doubly distilled from alkaline permanganate. The ammonia was redistilled over calcium hydroxide into cooled conductivity water. The mixtures were allowed to stand 1 hr. The density of particle distribution was determined for each mixture by the method of rhythmic counting. An 8-mm. objective and a 1OX ocular were used in all of the measurements. The results obtained are assembled in table 1. The data in table 1 show that the metaphosphate does not increase or stabilize the particles of colloidal dimensions, since the number of particles in the mixtures in which precipitation was entirely prevented did not differ significantly from that of the “blanks” on the original solution. It might be pointed out that the precipitation of calcium carbonate which occurred in the mixtures containing up to 1.2 p.p.m. of metaphosphate resulted from coprecipitation with the suspended impurities present initially in all of the mixtures. Photomicrographic evidence has shown that, in the presence of metaphosphate a t a concentration so low that the inhibition is not complete, the calcium carbonate crystals will form on any kind of suspended impurity that may be present in the solution. Under such circumstances many of the initially suspended par-

554

T. F. BUEHRER AND R. F. REITEMEIER

ticles might be deposited, with the obvious result that the ultramicroscopic count would be lowered. An independent experiment was designed to determine the “life” of a colloidal calcium carbonate particle if its existence were possible. For this purpose a sample of a calcium bicarbonate-ammonia mixture without metaphosphate was placed in the ultramicroscope cell and observed for a few minutes. It was found that the number of particles observable under these conditions increased to a maximum in about 5 min. and then began to decrease. This observation proves that the existence of such a

TABLE 1 Density of particle distribution i n the precipitation of calcium carbonate rhythmical counts with the ultramicroscope CONCENTRATION6 OF CONBTZTUENTU

Cat+

HCOa-

NHr

(NaP01h

p.p.m.

p.p.m.

p.p.m.

p.p.m.

50

155 5 560

50 50

50 50

50 50 50 50

50 50

155 155 155 155 155 155 155 155 155 155

560 560 560 560 560

560 560

560 560 560

5 4 3 2 1.5 1.2 0.9 0.6 0.3 0.0

T.F.B..

R.F.R.?

24 22 63 27 27 47 60 54 67 74 47 19 22 24 21

26; 15 14 80 38 33 54 79 44 61 51 39 24 24 16 20

a8 indicated

by

AYERAQB NUYBEB PER COUNT

0.21 0.18 0.72 0.31 0.50 0.70 0.49 0.64 0.63 0.43 0.22 0.21 0.20 0.21

* T.F. Buehrer. t R. F.Reitemeier. colloidal state is very transitory and that the particles grow rapidly to macroscopic dimensions. With respect to the theory of colloid stabilization the foregoing results indicate definitely the absence of any such stabilized colloidal calcium carbonate particles. If such a process occurred, it would enormously increase the colloidal content of the suspension, a condition easily detectable with the ultramicroscope. The very slight increase in the count may be accounted for in terms of the deposition of the precipitate on the suspended particles as rapidly as it is formed. It is of course possible that there might be particles too small to be visible even under the ultramicroscope. Even if the existence of such

PRECIPITATION OF CALCIUM CARBONATE.

I1

555

colloidal particles were demonstrable, it is still problematical whether hexametaphosphate could stabilize such particles. Gortikov and Malinovskaya ( 5 ) found calcite and aragonite to be positively charged. Hence the negative metaphosphate ion would coagulate rather than stabilize them. On the other hand, Chwala (1) has demonstrated the use of sodium pyrophosphate as a peptizing agent. The experimental evidence here presented has convinced the authors that the theory of colloid stabilization, to account for inhibition by metaphosphates, is not tenable. PHOTOMICROGRAPHIC STUDIES

ON

CALCIUM

CARBONATE

PRECIPITATED

UNDER DIFFERENT CONDITIOXS

It is recognized that the formation of crystal nuclei and the subsequent growth of the crystal are two distinct phenomena. The foregoing experiments demonstrated that metaphosphates do not influence the rate of nuclei formation. Hence it was thought that the inhibition manifested itself rn an abnormal or deranged crystallization process, in which either the growth of the crystal may be retarded or ’the form or habit of the crystals may be markedly changed. Preliminary microscopic examination of the firmly adherent crystals, above referred to, showed them to be very different in appearance from the typical calcite rhombs obtained under ordinary conditions of precipitation. The typical crystalline forms of calcium carbonate,-namely, calcite, aragonite, and p-calcium carbonate,-are well known (8). It is possible, however, for these modifications to assume a great variety of crystal habits (4). Kohlschutter and his coworkers (10, 11) have shown that the crystal habit of a substance may be markedly influenced by the presence of colloidally dispersed material in the solution in which the precipitation took place. They found, for example, that certain dyes, gelatin, agar, and other substances produced what they termed “somatoid” forms of calcium carbonate. France and his associates (2, 3, 9) demonstrated a similar effect in their work on the crystallization of copper sulfate, certain alums, and other highly soluble salts. A series of experiments was accordingly designed, which closely paralleled the inhibition experiments of the preceding paper (14) and in which changes in crystal form could be recorded in the form of photomicrographs. Procedicre for obtaining photomicrographs To mixtures containing 50 p.p.m. of calcium and 150 p.p.m. of bicarbonate, varying amounts of the respective phosphates were added, and an amount of ammonia solution subsequently added such that when the total volume was in each case brought to 100 ml. with distilled water, the final ammonia concentration would in all cases be 550 p.p.m. of ammonia. The mixtures were allowed to stand undisturbed for a definite period of

556

T. F. BUEHRER AND R. F. REITEMEIER

time, after which the supernatant liquid waa decanted to terminate the reaction. The crystals were then washed and, if loosely adherent, transferred to a microscope slide by means of a medicine dropper. I n cmes where the crystals adhere firmly to the beaker, and it seemed unwise to break them while removing them with a spatula, they were photographed directly in the bdttom of the beaker. A uniform magnification of 300 diameters was employed throughout. The photomicrographs so obtained will be presented in figures 1, 2, 4, 5 , 6, 7, 8, 9, and 10.

Calcium carbonate precipitates from the system Ca(HCOs)a-(NaPOs)s-NHa afhr 4 hr. (see jigure 1) The effect of the hexametaphosphate on the crystal habit is at once apparent. With no metaphosphate present, typical rhombs of calcite are obtained. In the presence of metaphosphate at concentrations of 0.3 to 0.6 p.p.m., the rhombs are fewer in number and larger in size, many of them showing various degrees of distortion. At 0.9 p.p.m. the original rhombic form has entirely disappeared, and only grotesquely deformed masses of somewhat smaller size rem& At 1.2 p.p.m., corresponding to the “threshold value” for this system as found in the preceding paper (14),only a few crystals were obtained, of about the size of normal rhombs but without definite form. Above this concentration no crystals whatever were obtained, Le., precipitation waa entirely prevented. These results are in striking accord with those found in the inhibitionexperiments, based on the analytical determination of the amount of calcium precipitated under like conditions. Calcium carbonate precipitates from the system Ca(HCO&-(NaPOS)s-NH, ajter 2 days (see jigure 2 ) Figure 2 illustrates the effect of time on the extent of deformation that may occur. The crystals are much larger in size, fewer in number, and much more distorted. In some the deformation has just begun to appear in the form of a roughening at the edges; in others there is complete distortion. The size of the normal rhombs, obtained in the absence of metaphosphate, does not change on standing. Beyond 1.5 p.p.m., no crystals whatever were obtained. Th’e change in size of the crystals precipitated from a solution of calcium bicarbonate and also from a natural water, as the metaphosphate concentration is increased, is shown graphically in figure 3. The natural water, included in the study because. of its unusual interest, came from a well near Goleta, California. Among many such waters studied by Rosenstein and his coworkers, the Goleta water proved to be exceptional in its behavior, requiring an extraordinarily high concentration of metaphosphate to inhibit precipitation. This behavior is probably due in part

I’PECIPITATION OF CALCIUM CAHUONATIS.

11

557

1

*.,

,

4

6

PRECIPITATION OF CALCIUM CARBONATE.

I1

559

to an unusually high concentration of calcium (216 p.p.m.) and magnesium (80 p.p.m.). I t will be observed that the mean diameter of the particle increases with the concentration of metaphosphate to a maximum a t a “threshold value” of about 1 p.p.m., thereafter decreasing sharply in the direction of a point where no precipitate whatever would be obtained. The size of the crystals appears to depend upon two factors: (1) the extent of the inhibition a t any given metaphosphate concentration, and (2) the actual supply of calcium and carbonate ions which are available for precipitation. There are apparently t u o steps in this process. The inhibition process requires that the crystals become progressively larger as the metaphos-

FIG.3. Mean diameter of calcium carbonate crystals as a function of metaphosphate concentrat ion in solutions from which they were precipitated.

phate concentration increases. This tendency is counteracted by a decrease in the supply of calcium and phosphate ions which is, in turn, controlled by the metaphosphate Concentration. In other words, by retarding precipitation the inhibition process causes larger crystals to form, and by eventually preventing precipitation it reduces the quantity of precipitate which can form.

Calcium carbonate precipitates f r o m the system Ca(HC03)2-(NaP03)6-XH3 after 82 days (see jigure 4 ) Figure 4 shows the enormous size and the extraordinary degree of distortion which thc crystals can attain on long standing. I t is significant to note that typical rhombs obtained in the absence of metaphosphate

560

T. P. BUEHNEH AND

I(.

P. BEITEMEfEn

remained urnchanged in form and were only slightly srnaller in average size at the end of this period. Rough measurements of the areas of the dktortcd particle in No. 2 and the average dimensions of those in No. 1 indicated a volume ratio of the order of 10,ooO:1, a change bmught about by a t:oncentrat.ion of only 0.9 p.p.m. of hexametaphosphate. The actual diamet,er of the divtori.ed crystal here shown was about 0.2 mm.

Calcium curbonale precipitates from lhe a&?

4 h,r.

q a t e m Ca(H(.!Oajr-Il;a,I:O, (see $gfisnre .if

~NK,

The effect of pyrophosphate is similar to that of the mi~taptiosphate, involving a progrer;sively iiiorcusing distortion (if thP cryst,& SI:^ figurc 5 ) . At 0.3 p.p.m., m e fiiids typical nodulnr gmutlis on the: fnctv of thi?

0

..

1 Fio. 1.Cxlciiirn carbonate precipitates from tho system Ca(IZCO,)T(NaPOl)sSlla aftor stiinding 32 days. 1, no (XsPOi)s; 2,0.9 p.p.rn. (NsPO&

rhombs. At 0.45 p.p.m., the original rhombic form has entirely disappeared.

Colczuni carbunale precipilules from the system C a ( H U O I ) * - . N ~ ~ f O , ~ ~ N H (see f i u r e 6 )

It was shown in rhe preceding paper (14) t.h& orthopliospliatos do not inliild the preripitai,ioo of calcium carhrtriat,~?in the same mnse ae do the currcs))miding meta and pyrtr salts. Tlic photomicrt~graphsof figure 6 substantiate this observation. A t a roiicent,ration of 0.3 p.p.m., the rhmlw arc still unchanged. At 0.9 p.p.m., which is approxiniately the threshold value for the other phosphates, there ix atill considerahle precipilat.ion, but it will he noted t,hat another species of crystal has made ita appearance, which is probably an orthophosphate of calcium.

560

T. F. BUEWKEH AND E. F. KEITEMEIEK

remained unchanged in fovm and were only slightly smaller in average size at the end of this period. Rough measurements of the are= of the distorted particle in No. 2 and the average dimensions of those in No. 1 indicated a volume ratio of the order of iO,WO:l, a changr brought about by a concent,rat.ionof only 0.9 p.p.m. of hexametaphosphate. The act,ual diamebrr of the distorted crystal here shown w m ahout 0.2 mm.

Calcium carbonate pvecipitates jrom the .qrtem Ca(11(:O$2 -IZ'a,P*07 ~NW, after 4 hr. (see fiure :7) The effcct of pyropbosphatc is similar t.o that, of the mt.tapiiosptiate, iiivol ving a progressively inwmsing distortion of t h e cry.stal,? r o i i illi.3p.p.m.

w.rljonatrr precipitation Figure 7 coirfirins this finding, since therc! is neither a distortion of the crystals rior an apyrecialdc changc in size up to the reiativcly high conceiit,ratiorr of 10 p.p.m. Calcium carbonale prenpitales from the qvstmL (h(HCOr)t ~Nal’Oa

(crystalline) in KOH-MI, 1t x crperirncnts illirstratcd in figures 7 atid 4 ~ v c ~sugg:astirl t: Ijg Kosc~,. “Our i~xpcrirnent,strn -hewn I h i ~ tthis (soltihie, stein (15), who wri r l

PIlECIPITATION OF CALCIUM CARIJONATE.

II

565

crystalline) trimetdphospiiate is ineffective as an inhibitor of calcium earlmiate preeipitation but hecomes just as effective as the hexamctaphosphate if it is first dissolved in dilut,e potassium hydroxide solution.” In the present instance tlie crystalline metaphosphate, hefore hning added to tlic calcium bicarhomate solution was madde alkeline to t h r extent of 2 2V with potassium hydroxide.

FIG.9. Calcium carbonate precipitates from the system Cs(HCOa).-(NaPO& in 2 .N KOH XH3 after 4 lir.

.Amount of (X~l’Oa)e:I, none; 2, 0 3 p.p.m.; 3, 0.6 p.p.m.

T l w photornicrographs show that im the presence of potwium hydroxide tlic cliaracteristie crystal abnormalities manifest themselve8. At a cnncoltration of thc metapBosphatP as 10s- as 0.5 p.p.m., t,he normal calcite rhombs are entirely ahsent, and in their places wc agairi find fewer, larger, and grossly diuturled crysials. T h e size reached a maximum at 1.0 p.p.m., and beyond 1.5 ;).p.m. no precipitate whatever wi*?fomed. These results suggest that the gotzwium hydroxi& may have acted to convert the

566

T. F. BUEHRER AND R. F. REITEMEIER

crystalline metaphosphate into the hexa-modification or to some active intermediate form. The question now arises: Would potassium hydroxide affect the ihhibiting efficiency of the hexametaphosphate? The results of a similar experiment to prove this point are shown in figure 9.

Calcium carbonate precipitates f r o m the system Ca(HC03)2-(SaP03)a in 2 N KOH-SH3 It appears that the base actually increased the efficiency of the hexametaphosphate (see figure 9). This result is the more striking because the high alkalinity would increase the concentration of carbonate ions and thus favor precipitation. The crystals show the same tendency to become distorted. The threshold concentration was found to be 0.8 p.p.m., as compared with 1.2 p.p.m. without the potassium hydroxide. It is possible that the base in some manner affects the polymerization of the hexametaphosphate to produce a more reactive phosphate. Calcium carbonate precipitates from the system Ca(HCOa)2-KaC1-NH3 (see fisure 10) Although as a rule calcium carbonate is more soluble in solutions of neutral salts than it is in pure water, it was expected that the extent of precipitation would be decreased, even in the absence of metaphosphate. The effect of a salt such as sodium chloride on the crystal habit is not readily predictable, although Marc and Wenk (12) found that sodium chloride and other simple neutral salts have a marked effect upon the formation of alum crystals from their saturated solutions. In the present experiment, where the sodium chloride concentration was carried through an extended range from 0 to about 12,000 p.p.m. (0.2 N ) , there is only a slight increase in the size of the crystals, while at the same time a progressive replacement of the calcite rhombs by the dumbbell-shaped aragonite crystals occurs. It is evident, therefore, that inhibition of the type which characterizes the metaphosphates does not occur even in. the presence of high concentrations of the ordinary neutral salts. It is merely a change from one allotropic modification to another, probably due in part to a change in the activity of the ions involved in the precipitation. X-RAP POWDER SPECTROGRAMS O F VARIOUS CARBONATES

The foregoing experiments raised the question as to whether the distorted crystals represented a fundamentally different crystal modification of calcium carbonate when precipitated in the presence of metaphosphate. To obtain evidence on this point, x-ray powder spectrograms of the precipitates, prepared in the presence as well as absence of metaphosphate, were obtained and these in turn were compared with the patterns yielded

568

T. F. BUEHRER AND R. F. REITEMEIER

by pure calcite and aragonite. The spectrograms were found in all cases to be identical with that of calcite. Thus, although the metaphosphate causes considerable distortion of the crystals, so that they exhibit no external signs of a definite structure, they are, nevertheless, true calcite. In natural waters, where magnesium may be present in considerable amounts, there is a tendency for magnesium carbonate to precipitate out also; under those conditions calcium carbonate often precipitates in the form of aragonite. RAT1O''OF

CALCIUM TO PHOSPHORUS IN THE CALCIUM CARBONATE PRECIPITATES

The foregoing photomicrographic studies have shown that the inhibitory effect of metaphosphates consists in part of a retardation of the crystallization of calcite, and, in addition, an extensive derangement of the normal growth of the crystal lattice. It seems necessary to assume that the metaphosphate has in some manner entered the lattice or has been adsorbed on the surface of the crystal. The latter view has been taken by Rosenstein (15) and by Hatch and Rice (7). It seemed of considerable interest to determine the order of magnitude of the ratio of Ca:P, and to ascertain how it changes as the metaphosphate concentration is increased. While the x-ray diffraction method indicates the nature of the crystals which were formed, it does not offer an answer to the question as to a possible combination of the metaphosphate with the calcium carbonate. To secure quantitative evidence on this point, four sets of precipitation mixtures were prepared, each set consisting of eight mixtures in which the concentration of metaphosphate was varied from 0 to 1.5 p.p.m. The concentration of calcium was 50 p.p.m., of bicarbonate 150 p.p.m., and of ammonia 567 p.p.m. These concentrations were the same in all of the mixtures, and the total volume was kept constant at 400 cc. The four series of mixtures were allowed to stand 3,21, 76, and 420 hr., respectively. At the end of these time intervals, the precipitates were filtered off, dissolved in hydrochloric acid, and warmed on the water bath for 10 hr. to hydrate the meta- to the ortho-phosphate. The solutions were then analyzed for calcium and phosphate, respectively. The results are shown graphically in figures 11, 12, and 13. Figure 11 shows the amounts of calcium precipitated as a function of metaphosphate concentration and of time of standing. It will be seen that the curves so obtained correspond closely to those of figure 4 of the preceding paper (14). Two significant relationships are evident from figure 12. First, the amount of phosphorus associated with the precipitated calcium carbonate increases with the concentration of metaphosphate up to a maximum which in all cases occurs at or near a concentration of 0.6 p.p.m., beyond

PREQXPITATIONOF CALCIUM CARBONATE.

11

569

which point the curves show a sharp drop. Below this concentration the curves have the characteristics of a typical adsorption isotherm. Apparently the metaphosphate is adsorbed by the precipitate in increasing amounts until this limiting concentration is reached, beyond which the amount adsorbed decreases. This decrease beyond the maximum is merely incidental, since the amount of precipitate that separates out decreases with increasing metaphosphate concentration. The amount of calcium precipitated decreases more rapidly than the amount of phosphate adsorbed increases, hence the molecular ratio of the two constituents decreases rapidly up to the metaphosphate concentration of 0.6 p.p.m.

FIQ. 11 FIQ. 12 FIG. 11. Calcium content of calcium carbonate precipitates FIG. 12. Phosphorus content of calcium carbonate precipitates

Another point of interest shown by figure 12 is the phenomenal increase in the amount of phosphorus adsorbed as the time of standing is increased. This is the more striking in view of the photomicrographic evidence, which indicated that, as the time of standing wae increased, the number of extraordinarily large particles as well as their size increased a t the expense of the smaller ones. Such a process should result in a markedly decreased surface and hence decreased adsorption. In figure 13 the atomic ratios of calcium to phosphorus are plotted against the metaphosphate concentration. The ratio decreases sharply and tends to level out to a constant value beyond a metaphosphate con-

570

T F. BUEHRER AND R. F. REITEMEIER

centration of 0.6 p.p.m. Because of the inherent limitations of the semimicro analytical method used, the curves do not converge to, or become precisely superimposable at, a particular value of the ratio. There is no apparent trend, however, since the points lie above and below a mean value of the ratio, which a t the “threshold inhibition value” of 1.0 p.p.m. was about 320. The deviations of the curves from this limiting value are within limits of error of the analytical methods. The magnitude of this ratio, in spite of its apparent constancy, does not justify the conclusion that a definite compound is formed between the metaphosphate and calcium carbonate. The phenomenon rather indicates adsorption on the surfaces of the calcium carbonate particles, which in turn is responsible for the subsequent abnormal growth of the

2”e

B

nm-

zmw-

!I-

\

i

!-

J

B5 -s --

‘It FIG.13. Molal ratio of calcium to phosphorus in calcium carbonate precipitates

crystal. The constancy of the ratio, which is remarkable in view of the extended period of time through which the experiment was carried, is probably to some extent fortuitous. The experiments do show, however, that the effect of the metaphosphate is not altogether an indirect one, as, for example, by affecting the activity of the reacting components in some manner, but a direct effect upon the calcium carbonate precipitate itself a t the time of its formation. DISCUSSION

The evidence presented in the preceding paper (14), as well as in the experiments just described, makes it possible to deduce a tentative mechanism for the inhibition process. It has been shown that the inhibition may be either partial or complete, depending upon the metaphosphate concentration. When the latter exceeds considerably the “threshold” concentration, there is a relatively permanent postponement of the pre-

PRECIPITATION OF CALCIUM CARBONATE.

I1

571

cipitation. Such an inhibition as well as complete prevention of precipitation takes place in the solution phase, notwithstanding the high degree of alkalinity on account of the free ammonia, together with a sufficient concentration of calcium and bicarbonate ions to favor precipitation. Such behavior cannot be justifiably attributed to complex formation between calcium and metaphosphate, in view of the extraordinarily small formula-weight ratio of metaphosphate to calcium, namely, 1:1000. This conclusion is also consistent with the fact that under such conditions the calcium is almost quantitatively titratable with soap solution. For the complete softening of boiler-feed water, on the other hand, Hall (6) finds that the corresponding Ca:NaPOa ratio must be a t least 1:4 at pH 8.5 and 1:7 a t pH 10. He attributes the softening action to the formation of a complex, the composition of which apparently varies with the pH value. Since in the present case complex formation occurs to a negligible extent, we must conclude that the action is largely an indirect one, involving either a stable electrostatic attraction between calcium and metaphwphate ion or a marked decrease in the activity of calcium ion due to the presence of the metaphosphate. Whatever the effect of the metaphosphate may be upon calcium ion in completely preventing precipitation of calcium carbonate, the experimental evidence regarding its inhibiting effect within the range of the threshold concentration of 1 p.p.m. points definitely to an interference in the growth of the calcium carbonate crystals. It was found that under these conditions the growth of the crystals was greatly retarded and the crystals formed were fewer in number, larger in size, and more or less distorted. The rate of deposition of the carbonate is approximately inversely proportional to the metaphosphate concentration. It is generally conceded that the formation of crystal nuclei and the subsequent growth of the crystal are entirely distinct phenomena. While certain investigators have been inclined to consider the effect of the metaphosphate to be one of stabilizing a supersaturated solution of calcium carbonate and thus preventing precipitation, the very fact that crystals do form argues against such a possibility. It was found, moreover, on the basis of ultramicroscopic observations that calcium carbonate does not exist in colloidal dispersion, or, if it does, its existence in that state is very transitory. Hence we are led to the conclusion that the metaphosphate does not interfere with the normally rapid transition from nuclear through colloidal to macroscopic dimensions, but that it retards the subsequent growth of the crystals. The experimental evidence indicates fairly definitely what occurs during this retarded process of crystal growth. First, it is found by analysis that the metaphosphate actually enters the lattice or is adsorbed on the surfaces

572

T. F. BUEHRER AND R. F. REITFMEIER

of the crystals. The greatest amount of phosphorus waa found in those crystals which exhibited the greatest degree of distortion. Secondly, the abnormal growths usually make their appearance on the faces and apices of the normal calcite rhombs first formed. This finding is in accord with the view of Pauliig (13),who states that such an effect will usually manifest itself first at the apices of a crystal lattice, which are the points of greatest instability and reactivity. Evidently the metaphosphate is adsorbed by the crystals and thereafter the normal procem of crystal growth is deranged. Similar abnormalities in the process of crystallization as well aa in the form of the crystal obtained have been encountered by other workers. Marc and Wenk (12) were among the first to observe the effect of small amounts of a foreign adsorbable substance in preventing the crystallization of potassium sulfate from its saturated solution even after long standing. They showed, further, that such foreign substances did not prevent crystals from dissolving in their own unsaturated solution. In the light of their findinge it seems reasonable to conclude that any nuclei or submicroscopic crystals of calcium carbonate are readily dissolved because they have a higher solubility than the large crystals, and because their solubility is not influenced b y the metaphosphate. The larger crystals thus grow at the expense of the small ones, but their growth is retarded and their shape distorted, owing to the presence of the metaphosphate. Keenen and France (9) explain the effect of gelatin and certain dyes on the rate of growth of potassium alum crystals by assuming that the foreign substance is adsorbed by the crystals ‘ I . . . i n areas of different concentration, corresponding to the different fields of force existing at the various crystal faces. Those faces having alternate planes of like ions exhibit much greater adsorbing power than those with a checkerboard arrangement of positive and negative ions.” Such a mechanism may serve to explain the exaggerated growths from certain faces of the calcium carbonate crystals as shown in figure 2 (2 and 3). This behavior is also consistent with the nature of the “etch-figures” used by the mineralogist in determining the symmetry of a crystal. In this case the action of the solvent begins at certain points and proceeds with greater rapidity in certain directions than in others; in addition, it may produce different figures on certain faces and leave others entirely unaffected. A strikingly similar type of abnormality in the case of calcium carbonate crystals was found by Kohlschutter and Egg (10, 11). They obtained such distorted crystals, which they called “somatoids,” in the presence of various colloidal sols and dye solutions. They concluded that such somatoid forms result from the adsorption of foreign substances present in solution upon the crystals of calcium carbonate during the period of their precipitation.

PRECIPITATION OB CALCIUM CARBONATE.

I1

573

I n the present case we have an example of derangement of crystal growth by reason of the adsorption of a substance which is not colloidally dispersed but in true solution, and, furthermore, is present in an exceedingly small concentration. These facts, together with the observation that inorganic salts of other elements, similar to sodium metaphosphate, and other simple salts like sodium chloride, even when prkent at high concentration, are entirely without such inhibiting action, indicate that the action of sodium metaphosphate is highly specific. SUMMARY

1. Ultramicroscopic studies have been made on solutions in which calcium carbonate was formed by precipitation in the presence of progressively increasing concentrations of sodium hexametaphosphate. The results indicate conclusively that calcium carbonate does not go into the colloidal state nor is such a state stabilized by the presence of the metaphosphate. 2. Photomicrographic studies indicate that the metaphosphate promotes the formation of larger, fewer, and grossly distorted crystals in place of the typical rhombs of calcite. It was found that the meta-, p y r e and ox tho-phosphates possess in varying degrees this ability to hinder the process of crystallization. 3. The soluble crystalline sodium metaphosphate does not possess the property of inhibiting the precipitation of calcium carbonate, but acquires it in the presence of potassium hydroxide. 4. Neutral salts such 89 sodium chloride promote the transformation of calcite to aragonite rather than to the abnormal somatoid forms. This may be a simple solubility effect. 5. X-ray powder spectrograms have shown that the distorted crystals obtained in the presence of these phosphates are true calcite, although their exterior surfaces may show complete absence of typical polyhedral faces. 6. Chemical studies of the precipitates indicate that the metaphosphate is adsorbed by the calcium carbonate in amounts such that the Ca:P ratio attains a constant value of approximately 300 in the presence of metaphosphate concentrations greater than 0.6 p.p.m., which is the concentration a t which the adsorption reaches a maximum. 7. The mechanism of the inhibition process is conceived to involve primarily a restricted or deranged crystallization due to the adsorption of the metaphosphate on the crystal faces.

The authors desire to acknowledge with thanks the assistance and encouragement of the Shell Chemical Company, which sponsored this research, and the kindness of the General Electric x-Ray Corporation of

574

RAY CALVIN CHANDLER

Chicago, Illinois, and of Dr. D. H. Reynolds in particular, in making the powder spectrograms of the various carbonate and phosphate samples obtained in this study. REFERENCES

(1) CHWALA, A.: Kolloid-Beihefte 31, 222 (1930). (2) ECKERT,T. S.,AND FRANCE, W. G.: J. Phys. Chem. 31, 877-81 (1927). (3) ECKERT, T. S.,AND FRANCE, W. G.: J. Am. Ceram. Soc. 10, 579-91 (1927). (4) GOLDSCHMIDT, A.: Atlas der Kristallformen, Volume 11, p. 5. Carl Winters, Heidelberg (1913). (5) GORTIKOV, V. M., A N D MALINOVSKAYA, N. P.: Colloid J. (U. S. S. R.)2, 42933 (1936). (6) HALL,R. E.: U. 5. patent 1,956,515(April 24, 1934), p. 5. (7) HATCH,G. B., AND RICE, OWEN:Ind. Eng. Chem. 31, 51-7 (1939). (8)JOHNSTON, J., MERWIN,H. E., AND WILLIAMSON, E. D.: Am. J. Sci. [4] 41, 473-512 (1916). (9) KEENEN,P. G., AND FRANCE, W. G.: J. Am. Ceram. Soc. 10, 821-7 (1927). 10) KOHLSCH~TTER, V., A N D EGG,C.: Helv. Chim. Acta 8, 470-90 (1925). 11) KOHLSCH~TTER, V., AND EGG,C.: Helv. Chim. Acta 8, 697-703 (1925). 12) MARC,R.,AND WENK,W.: Z. physik. Chem. 88, 104 (1909). 13) PAULING, L.: Personal communication. 14) REITEMEIER, R. F., AND BUEHRER, T. F.: J. Phys. Chem. 44,535 (1940). 15) ROSENSTEIN, L.: Private communication.

A SENSITIVE STATIC VAPOR PRESSURE APPARATUS' RAY CALVIN CHANDLER* Department of Plant Nutrition, University of California, Berkeley, California Received June 67,1939

The vapor of an aqueous solution under its own vapor pressure in a closed system approximates the behavior of an ideal gas. Such a system is constantly in equilibrium, capable' of rapid reversible adjustments to variation in temperature, pressure, and concentration. A satisfactory static vapor pressure apparatus is very desirable for the calculation of thermodynamic functions in the study of solutions. Hitherto the use of this method has been limited by the difficulties involved in the preparation of experimental solutions which were sufficiently free from residual gases 1 This paper constitutes a portion of a thesis submitted by the author to the Graduate Division of the University of California in partial fulfillment of the requirements for the degree of Doctor of Philosophy, December, 1937. * Present address : Department of Botany, University of Arizona, Tucson, Arizona.