American Mineralogist, Volume 59, pages 947-963, 1974

Gontrolled Nucleation and Crystal Growth of Various CaCOuPhases by the Silica Gel Technique

Jnurs W. McCaurEyl, AND Rusruu Roy

Department ol Geochemistry and Mineralogy and the Materials ResearchLaboratory, The Pennsyluania State Uniaersity,

Uniaersity Park, Pennsyluania 16802

Abstract

The effect of the environment on the precipitation and crystal growth of CaCOe fromaqueous solutions at ordinary temperature and pressure has been investigated using the silicagel technique. Thermodynamics fails to predict many of the observed assemblages, suggestingthat kinetic mechanisms are more important for these quasi-metastable occurrences. Theeffect of concentration of reactants, pH, and impurity ions on the precipitation of CaCO' ina silica gel has been determined. Aragonite formation, in the absence of impurity cations,seems to be caused by the entrapment of HCO;, which is a function of the relative concen-tration of COr- to HCOi. Synthesis data for vaterite suggests that this phase may be relatedto the precursor formation of CaCO".H,O and also that it is not a stoichiometric CaCOsphase; rather, a Ca-rich material with small incorporations of (OH), HCO,, or CO" (aq.).The morphology of calcite and aragonite, in the absence of impurity cations, is a pronouncedfunction of pH. In the presence of the impurity cations, Sf*, Mg'*, and Ni'*, aragonite forma-tion is controlled by its epitaxial growth on aragonite-like nuclei; for Sr"* unequivocal evidenceindicates that this nucleus is SrCO., whereas, for Mg the nucleus seems to be MgCO"'3H,O,nesquehonite.

The partition of these impurity cations between solution and the various precipitatingsolids is a complex function of pH, relative concentration of reactants and impurity ions, andamounts of various CaCO' phases precipitating simultaneously. The partition coefficients forSr in aragonite and calcite were 4.37 and 1.04 respectively. Single crystals of calcite, with 7.5mole percent Mg, were grown by this technique.

Introduction

Calcium carbonate precipitation from aqueoussolution is affected in varying degree by the environ-ment, but the mechanisms are still somewhat ob-scure and unverified by direct evidence. Five poly-morphic modifications of crystalline anhydrousCaCO3, each with a fairly well defined temperatureand pressure field of stability, have been described:calcite I (hereinafter called calcite), calcite II, cal-cite III, aragonite, and vaterite. However, calcite,aragonite, and vaterite may be prepared either singlyor simultaneously by precipitation from an aqueoussolution at ordinary temperature and pressure con-ditions, even though only calcite is thermodynam-ically stable. The metastable formation of vateriteand aragonite at roo.m conditions results from inter-ferences in the equilibrium process by kinetic con-

l Now at the Army Materials and Mechanics ResearchCenter, Watertown, Massachusetts 02172.

947

siderations. Many reviews of this problem havebeen published: Goto ( 1961 ) , Fisher (1962) ,Bricker and Garrels (1967), etc.

Most of the investigations in this area have beenconcerned with the effect of impurity ions on, thephase that separates from solution. Johnston, Mer-win, and Will iamson (1916), in an early classicwork, showed that when Pb was added to theprecipitating medium aragonite formed; these crys-tals had an optically distinct nucleus, which ap-peared to be PbCOB, suggesting epitaxial control.Since then, other workers (Murray, 19541' Wrayand Daniels, 1957; Zeller and Wray, 1956; Kitano,1962, 1964; Kitano and Hood, 1962; Simkiss, 1964;Bischoff and Fyfe, 1968) have shown how othercations such as Sr, Mg, and Ba affect calcite andaragonite formation. MacDonald (1956) calculatedthermodynamically that at least 30 mole percentof a typical impurity ion (e.g., Sr) in crystallinesolution is needed to stabilize the formation of

948

aragonite in the calcite stability field. Using a uniqueconcept of adjusting the dielectric constant of variousCaCOs precipitation environments, Goto (1961)analyzed the question in terms of hydration energiesof the ions. His results, although quite interesting,seem inconsistent with those of other workers.

The influence of Sr and Ba should be very similarto that of Pb as described by Johnston, Merwinand Williamson (1916), and this has been con-firmed in general. Mg, however, also favors aragoniteformation, although somewhat anomalously. Lipp-mann (1960) proposed that Mg poisons the grow-ing calcite crystals, thus allowing for the eventualsupersaturation and precipitation of aragonite'Bischoff (1968a) also has established the tendencyfor Mg to poison growth. Buerger (1971) hassuggested that Sr can influence the precipitation ofaragonite over calcite, by providing nuclei with anaragonite structure.

In general, analyses of natural CaCO3 poly-morphs for trace elements (Siegel, 1961; Stehli andHower, 1961) and laboratory investigations (Ox-burgh, Segnit, and Holland, 1959; IJolland et al,1963, 1964) indicate that calcite is generally higherin Mg and lower in Sr, Ba, and Pb than is aragonite.

Aragonite formation also seems to be favoredby high pH and high water temperature (greater than70"C). But here again, few hypotheses concerningthe mechanisms involved have been ofiered.

Vaterite has always been a special problem, inthat it has a vastly different crystal structure thancalcite and aragonite (Meyer, 1969), and becauseuntil recently it had no known T-P stability region.Its natural (McConnell, 1960) and laboratory oc-currence has always been quite intriguing. Theformation of vaterite is promoted by the presenceof barium(Kitano,1962; Goto, 1961) and by watertemperatures of 40o-70'C (Kitano and Hood,1962; Johnston, Merwin, and Williamson, 1916;and Dekeyser and Degueldre, 1950). Dekeyser andDegueldre (1950) have also shown that almost100 percent vaterite is formed when CaCl2 is addedto Na2COs, but when NazCOr is added to CaClz,almost 100 percent calcite precipitates.

Classical low temperature experimental techniqueshave shed much light on these problems, but havealso created some new ones. In these techniques,reaction and precipitation normally proceed veryrapidly, usually with formation of crystals much toosmall for individual study. Preliminary experimenta-tion using a silica gel technique (Vand. Henisch, and

I. W. McCAULEY, AND R. ROY

McCauley, 1963) for the crystal growth of varrous

slightly soluble materials, including CaCO3, indi-cated that this technique would be particularlyadaptable to CaCO3 problems. During this initialwork, our first attempt to grow crystals of CaCOsresulted in the formation of all three CaCO3 poly-morphs side by side in the test tube. Spurred onby these preliminary results, a detailed study wasinitiated to examine the synthesis of CaCO3 incontrolled environments (McCauley, 1965). AfterMcCauley and Roy (1965, t966a, b) describedthe silica gel crystal growth of CaCOg and SrSO+,several other investigators re-examined the crystalgrowth of CaCOg in gels (Nickel and Henisch, 1969;Barta and Zemlicka. 1971; Schwartz et al' l97l)'

An extensive literature survey revealed that Fisherand Simons (1962a, b) and Morse and Donnay(1931) were the first to grow crystals of CaCOsin a gel.

These initial studies demonstrated that fairly largecrystals may be routinely prepared, and then sepa-rated and studied independently by various tech-niques. Even more intriguing is that the crystalsmay be observed during their nucleation and growtheither macroscopically or microscopically since thegel is optically transparent. Further, the pH and theconcentration of reactants and impurity ions canalso be independently adjusted. Moreover, extran-eous effects on the precipitates are minimized sinceearly formed nuclei are delicately held in the posi-

tion of their formation, separate from all otherfirst formed nuclei. This minimizes effects due toprecipitate-precipitate interaction and crystal im-pact on the bottom of the container. The dataobtained, then, should yield information concerningthe mechanlsrns which operate to alter the predi-

cated equilibrium assemblage of phases and theoccurrence, both natural and synthetic, of the var-

ious CaCOr polymorphs.

ExPerimental Method

The use of gels as .a reaction medium has re-

ceived sporadic attention since 1896 when R' E'

Liesegang first observed the periodic precipitationof slightly soluble salts in gelatin. "Liesegang Rings"

inspired many chemists and mineralogists to study

othir reactions in various colloids; among the early

workers was Holmes (1917), who successfully grew

fairly large crystals of a variety of materials'The technique can be described as crystal growth

bv reaction: two solutions of soluble salts are

GROWTH OF CaCOs PHASES BY THE SILICA GEL TECHNIOUE 949

4 0 5.O 6.0 7.O 8.O 9.O tO.O I t.O tz.O | 3.O

p H +

Frc. 1. Titration curve a;,tlr$.O.t5 molal Na^SiO,:9H,O

brought together by diffusion through a gel withsubsequent nucleation and crystal growth takingplace by precipitation of a reaction-formed super-saturated phase within the gel. Various mechanismshave been suggested to describe the function ofthe gel in the crystal gowth process. fn general,however, the gel limits the number of critical-sizenuclei that are formed and decreases the rate ofgrowth by controlling diffusion of the reacting ionsand, more importantly, by governing the speed ofreaction at the crystal's growing surface. The gelis assumed to be chemically inert to any reactionthat takes place and may be thought of as analogousto a glass sponge. For more details and discussionsof these mechanisms see McCauley (1965) andHenisch (1970).

A gel may be fo.rmed using a wide variety oftechniques and materials; in this work silicic acidgel was prepared by the acidification of aqueoussolutions of Na2SiO3.9HrO (Fisher Scientific cer-tified reagents of 99.93 percent purity as determinedby N. H. Suhr of the Mineral Constitution Labora-tory). See Iler (1955) for extensive discussionsof colloidal silica. Figure 1 shows the pH of acetic-acid-silicate solutions as a function of milliliters ofacid added to 10 ml of a constantly agitated silicate

sillution. The pH of the gel is measured immediatelybefore gelation and, therefore, does not refer to thepH of the semisolid gel material or of the interstitialsolution. The quoted pH is assumed to be a closeapproximation of the average hydrogen ion con-centration in the environment of the growing crystals.

The pH of the solutions was determined by aPhotovolt pH meter model 115 with a Beckmanfiber junction reference electrode (calomel internals)and a Beckman glass electrode (silver-silver chlorideinternals) capable of measurements within 0.05 pHunit. The arrangement was calibrated in the nor-mal way using various buffer solutions of knownpH. Acetic acid was used to gel the sodium silicatesolutions because the pH of a silicic acid gel willnot change with time only if a weak acid is used(Hurd and Griffith, 1935).

There are several ways, as shown in Figure 2,in which crystals can be grown in a gelatinous me-dium. The simplest method is to incorporate asolution of one reactant salt in the acid-silicatemixture before gelation occurs and to pour a solu-

REACToN {AREA t

r 5 .o

t 4 5

r 3 . o

m l . t 2 . 5

I N H A ct 2 0

I t .5

I t .o

to .5

tO;O

9 . 5

9.O

METHOD I( o )

METHOD Itr( c )

REACTANT

GEL WITHREACTANTINCORPORATED

CONTAINER OFSOLID OR L IQUIOREACTANT

REACTION AREA

METHOD tr( b )

t)" " l '

METHOD E( d )

Frc. 2. Some methods for growing crystals in a gel.

I .14. McCAULEY, AND R. ROY

tion of the second salt on the gel after gelation (IV).

Another method is to form a gel over a packet(e.g., filter paper) of one of the reactants and pouron a solution of the second salt after gelation oc-curs; reactant concentration effects can only beapproximated in this case (I). By a U-tube method(II), both reactant concentrations are well defined

T.qsle 1. Some Phases That Have Been Grown by theGel Technique

C.v i la ls Method i Reactanc S ize {mm) Remarks

and reaction point concentrations can be approxi-

mated. The two-gel technique (III) involves pre-

paring a gel containing one of the reactants in the

bottom of a test tube as in method IV. However,

a second gel, called the reaction gel, is formed over

the first gel and a solution containing the second

salt is then poured over the reaction gel.

Calcium carbonate can be prepared by any num-

ber of different chemical reactions. The following

reaction was utilized in this study:

NazCO, * CaCl, --+ CaCOe * 2NaCl.

Since sodium is part of the gel, it is advantageous

to have sodium as part of one of the reactants in

order to minimize the amount of foreign cations

present. All the chemicals are Fisher certified reagent

grade.The concentrations of the reactants at the point

of reaction were calculated and approximated by

methods which depended on the method used and

which assumed classical diffusion laws. However,

the exact concentration ig impossible to know ex-

cept by direct measurement, and the data should

be taken to reflect relatiue changes in concentration

of the reactants, rather than absolute values of

concentration. Impurity cations (Sr, Mg, Ni, and

Ba) were added directly to the sodium silicate-acidsystem before gelation in predetermined concentra-

tions. Hence, the concentrations of these cations

are much more accurately known than the absolute

reactant concentrations.The various phases were identified by micro-

scopic observation and X-ray diffraction analysis.

Crystals were separated from the gel by vigorous

agitation of the final gel and crystal system in a

dilute Ca(OH)2 solution. This was carried out

repeatedly until all of the gel was removed from

the crystals. Crystals to be analyzed by emission

spectroscopy (by N. H. Suhr and D. J' Rhine of

the Mineral Constitution Laboratory) were further

washed alternatively in distilled water and solutions

of sodium hydroxide. The spatial distribution of

these impurity cations in some of the crystalline

products was determined by the use of an Anr,

electron microprobe operated by E. W. White and

G. R. Zechman, Jr., both of the Mineral Constitu-

tion Laboratory.Table 1 lists some of the more important crystals

synthesized during the course of this study' Note

especially that carbonates, sulfates (McCauley and

Roy, 1965), phosphates, chromates, and tungstates

CnSO4 2H20

BaSO4

PbSOa

SrSO4

PbCrOa

Cu3lPOa)2 3H20

C:HPOa 2H20

ca5{Po4)30H ?

CaF2

SrF2

L4sFz

BaCIF

CaWO4

GCO3

CaCO3 HzO

BaCO3

SrCO3

Nd2(CO3)3 8H2O

Cu2O

CuCl

Hg l . H9212

Pbl2

Sulfares

l V , C i C l 2 + H 2 S O 4 5 x 0 5

l l , CaCl2 + CuSOa 30 x I

l l l , C a C l 2 + H 2 S O a 5 x 0 5

l l , B a C l 2 + C U S O 4 l x l

l l , Pb {C2H3O2)2 + CuSOa 1 x I

l l . S r C l 2 + C u S O 4 l x l

Chromates

l l , Pb {CzH3Oz}2 + K2CrOa I x 1

Phosphates

lV , CUSOa + H3PO4

l V , C a C l 2 + H 3 P O 4 2 x 3

lV , CaCl2 + H3O4

Fluor ides

l r . c a c l 2 + N a F < 0 0 1

l l , s .c r2 + NaF <0 01

l l , M s C l 2 + N a F < 0 0 1

l l , B a c l z + N a F 2 x 2

TunStates

l l , C a c l z + N a W O a l x l

Carbnates

A l l , G C l 2 + N a 2 C O 3 l x l

l V , C a C l 2 + N a 2 C O 3 + s u 9 r l x l

l l , l V , B a C l 2 + N a 2 C O 3 1 x 0 l

l l . l V . S r c l z + N a 2 C O 3 1 x 0 l

l l f , N d c r 3 + N a 2 c o 3 2 t 1 , 0 1

Oxrdes

i l t , t v , c 0 s o a + N H 2 o H H c l 0 5 x 0 5

Chlor ides

I l l , CUSOa + NH2OH HCI I x l

tod ides

l V , H g C l 2 + K t O 2 x O 2

l V , P b ( C z H g O z ) : + K l 4 x 0 5

Metak

l l l , C U S O 4 + N H 2 O H H C I 2 x 0 5

Al l , Pb(C2H3O2)2 + Zn 0 I

H igh Ca reg g ismat ic c rys

High SOa reg acrcular crys

60"c

Usd ludox AS for gel

Spherulites and

liesegang tings

Sing le c ry*a ls ; b lade l i ke

L iesegang r ings & sph€ru l ; t * .

& radial crys clusters al

60 'c

Spheru l i t i c

Single crystats

Cl€ar spherulites

Ca lc i re , a ragon i te & va ter i t c

c .y i la ls & spheru l i tes

Single cryilals

Sheef l i ke c ry*a ls and

Sheef l i ke c rys ta ls and

Lavender p la tes and

Dark red to rransparentye l low cub ic c ry r ta ls

Ciear to dark green cryilak

rhat decompose in

L ieseganq r in$ o f c ry {a ls

tha t migra te down lub

Need le and 3d growth ,

ep i ta ! ia l l y on Cu20

Smal l need les & p la tes

growing on the end of

GROWTH OF CaCOa PHASES BY THE SILICA GEL TECHNIQAE 951

(McCauley and Gehrhardt, 1970), as well as oxidesand metals (McCauley, Roy, and Freund, 1969)may be synthesized; the latter by appropriatelycontrolling the oxidation potential of the gel. Itseems that by appropriate manipulation of thecrystallization environment (Eh, pH, complexingagents, impurity ions, concentration of reactants,temperature, etc.) a wide variety of slightly solublematerials may be successfully synthesized into fairlylarge (0.5 to 1 cm) single crystal size.

Experimental Results

I. Influence a'f pH qnd Concentation of Reactants

A. Phase Distribution

Both the pH of the reaction area and the relativeand absolute concentration of the reactants have apronounced effect on the relative percentage ofCaCOg phases that are formed; Figures 3-5 il-lustrate these results. Using method I (Figure 3)the concentration of CaCls was kept constant atabout 0.45 molar while varying the molar ratio

CaClz:NazCOs from 14.5 to 1.45. Primarily tosubstantiate the method I results, two series of runswere employed using method III; NazCO3 con-centration was kept constant at 0.09 molar in oneseries (Fig. 4) and 0.36 molar (Fig. 5) in another,while varying the CaCtr2 concentration from 0.65to 0.07 molar. The relative percentages indicatedin the figures were determined by relatively simplecounting techniques under a low power binocularmicroscope; the percentages (l 1O percent) repre-sent numbers of individual crystals, regardless ofsize or total weight.

Both aragonite and vaterite formation seem to bea pronounced function of the pH of the reactionarea in the gel and, less directly, of the concentra-tion of the two reactants. In particular, pH valuesroughly between 8 and 10 seem to be most favor-able for the precipitation of aragonite and vaterite;reactant concentration effects are observed but un-resolved. Further, in all of the test tube runs arugo-nite crystals formed primarily in regions of the gelcloser to the carbonate solution. while vaterite

iz s v . l

{

o / o r s o f

P h o s e s

A R A G O N IV A T E R I T

C A L C I T E

A R A G O N I T E

7 8 p 6 * 9 l o

e ffiffiffi.=r+.s7 8 p H + 9 l 0

.ffiff i.=s.ze

7 8 p H - 9 l O l l

o 94!,!!!bro* = r.+s

T2 5 % |

I

o/ " ' s o f

VATE R IT

A R A G O N I T E

P H + s

A R A G O N I T E

'3:ii#f;8i';="'Frc. 3. Percentage of phases formed as a function of reactant concentration and gel pH-

using Method [.

952 I. W. McCAULEY, AND R. ROY

V A T E R I T E

A R A G O N I T E

C A L C I T E

9.0 to 10.5

above 10.5

V A T E R I T E

A R A G O N I T E

C A L C I T E

I251" I

" /o 's of

P h o s e s

1,57" I

o / o ' s o f

P h o s e s

z 8 p t * g l o

affiff i=z.zz

V A T E R I T E

A R A G O N I T E

C A L C I T E

7 a . . 9 l 0p h +

O . 3 4 m C o C l rB

6 : 6 - 9 , N . 2 C O 3 =

" ' t o

6 7 8 p g * g

. O . l S m C o C l z - t . O' o . o 9 m N o 2 c o 3

preferred areas closer to the calcium solution. Inaddition, high initidl concentrations of both reactantsresulted in the formation of a pronounced band oftightly clustered crystals of rapidly grown calcitecrystals; smaller concentrations resulted in nodiscernible band. Figure 6 illustrates these observa-tions on two runs carried out at the same pH.Complete mixing of reactants throughout the gelis restricted in the fqrmer case and can substan-tially modify the interpretations of the results.

For one sequence of runs when the concentra-tion of both reactants'was high (Fig. 5a), vateriteprecipitated in appreciable amounts at a pH ofaround 8.00, confirming again that vaterite fcirma-tion favors high (rplative) concentrations of Ca.Additional, but incomplete, studies indicate thatvaterite is the predpminant phase formed in gelsof pH less than 6.0. Schwartz et al (1971) hAvealso made observatlons on pH dependency of thegrowth of vaterite.

B. Morphology ol CaCOt

Calcite formed in various rhombic-type singlecrystal and polycrystalline aggregates over the whole

o O . O 7 2 m C o C l z - O . a OO . 0 9 m N 0 2 C O 3

pH range investigated. Vaterite and aragonite oc-curred in various crystal clusters and spheruliticmorphologies. All morphologies were identified byX-ray powder diffraction. Three distinct morpho-logic types of calcite were identified:

pH Calcite Type

7 to 9.0 or 9.5 Various single crystal rhombictypes;

Frc. 4. Percentage of phases formed as a function of reactant concentration and gel pH-using Method III.

feathery-polycrystalline ag-gregate with rhombic outline;spicular or spherultic type.

Two types of calcite rhombs are illustrated in Figure7A: a normal rhomb and a corner growth (hopper-like growth) rhomb resulting from rapid crystalgrowth. Figures 7B and 7C depict other crystals ofcalcite and aragonite grown at near neutral pHvalues. Poisoning of single crystal calcite rhombstakes place in higher pH solutions and results inthe formation of feathery and of spicular varieties.The feathery type, a polycrystalline aggregate witha rhombic outline (Figs. 7D and 8G), has a rhombo-hedral core or nucleus upon which a mass of finely

C A L C I T E

V A T E R I T E

C A L C I T E

GROWTH OF CaCO" PHASES BY THE SILICA GEL TECHNIOUE 953

2 5 " h

o / o ' s o f

P h o s e s

V A T E R I T E

A B A G O N I T E

c A L C I T E

8 p H + 9

2 5 " h

" / o ' s o l

P h o s e s

6 7 8 p H * g l O

. 0 . 1 8 m C o C l 2 - ^ ^ ^- O . 3 6 m N 0 2 C 0 3

Frc. 5. Percentage of phases formed as ausrng

acicular crystals nucleate. The rhombic outline isprobably preserved because of the symmetricalovergrowth onto the initially formed calcite rhombo-hedron. Increased poisoning in more basic gels(above 10.5) results in the formation of the spiculartype (Figs. 7E and 7F). This particular variety wasinitially confused with aragonite due to the acicularmorphology. Figure 7G and 7H show, respectively,the beginning and intermediate stages of the poison-ing of a calcite rhomb which would lead to thecrystallization of either the feathery or spicular type.

Figure 8 depicts various bbqerved crystal clustersand spheruliticlike morphologies of aragonite andvaterite; neither precipitated as isolated single crys-tals. Again more nearly single crystalline materialformed at low pH values, (Figs. 7B andTC), whereasin more basic gels spicular-like and dense, finelypolycrystalline masses occur (Figs. 88 to 8G and 7F).Figure 8H shows vaterite spherulites formed underthe conditions diagrammed in Figure 5A; a botry-oidal-like vaterite (not clear spherulites) occur inmost runs very near to the CaCl2 solution. Both

V A T E R I T E

A R A G O N I T E

C A L C I T E

'ffiffi=o.s+s

7 a p H - 9 l o l l

o 9'97=2 t,,co c'=t . o.zo- 0 . 3 6 m N o 2 C O 3

- '

reactant cohCentration and gel pH-

7 I p H - g l O

offiffir=r.e'

function ofMethod III.

C A L C I T E

A R A G O N I T E -N OD I S C E R N I B L EF I R S TR E A C T I O NA R E A

C o cS A C K

R U N N O . I 5 0

0 . 4 5 m C o C l 2

O 3 l m N o r C O .

( o )

R U N N O . I 6 2

O . 4 5 m C o C 1 2

O . O 7 8 m N o z C 0 3

( b )

C A L C I T E

V A T E R I T E

x x

X x x

xx

X X

Frc. 6. Spatial distribution of crystals in the gel.

954 t. W. McCAULEY, AND R. ROY

FIc. 7. Various morpholosies of calcite: (A) pH : 6.6, (B) pH : 7.3, (C) pH : 7.3, (D) pH : 9.8, (E) pH : 10.5, (F) pH :10.6, (G) pH : 10.2, (H) pH : 8.9 (CC : calcite, A : aragonite).

the vaterite in Figure 8H and the aragonite in Figure7F are oolitic in appearance, suggesting a commonmechanism.

C. Discussion ol Results:

McConnell (1960) has suggested that calciumhydroxide catalyzes the formation of vaterite. Thedoubled lattice parameters of Ca(OH)z (2a -7.L9 A, 2c = 9.8 A) are quite similar to thoseof vaterite (pseudo-cell) (a = 7.16 A, c (pseudo)= 8.5 A, c = t6.94 A) implying that epitaxy ofvaterite on initially formed Ca(OH)2 could beimportant. However, careful X-ray powder diftrac-tion analysis of hand picked vaterite spherulitesrevealed that only CaCOg.H2O seemed to be in-corporated within the spherulites. The lattice param-eters (a = 6.09 A, c = 7.53 A, Kohatsu andMcCauley, 1973) and powder patterns of CaCOe.H2O are very similar to vaterite. Hence, it appearsthat epitaxial control of vaterite formation byCaCOg.HzO may be more important than pre-cursor Ca(OH)p precipitation.

A recent crystal structure analysis of CaCOs .HzO(Kohatsu and McCauley, 1973) showed that itsstructure is intermediate between those of calciteand aragonite on the one hand, and vaterite on the

other. Meyer (1969), in a detailed crystal structureanalysis of vaterite, clarified its relationship to thebaestnaesite (RFCO3) series of minerals and con-firmed the irregularity of the O-O interatomic dis-tances in the COr'- group. The similarity of theirregular O-O distances to those in CaCOa'HzOand NaHCO3 (Sharma, 1965) strongly suggests thepresence of undetected hydrogen bonding in thevaterite structure. These structure data in con-junction with the fabrication criteria imply thatvaterite may not be stoichiometric CaCOs, butrather a calcium rich material (Ca > CO3), withsmall incorporations of (OH)-, HCO-3 or CO2.One other possibility is that some Na* substitutesfor Ca'*, allowing for the substitution of (OH)-

or (HCOg)-.

Figure 9 is a somewhat schematic representationof the relative concentration of various carbonatespecies and the solubility of CaCOe as a function ofpH; in a strict sense this figure is valid only for verydilute solutions. Freiser and Fernando (1963) showhow effective equilibrium constants vary as a func-tion of ionic strength:

N/tp K i ' : p K r -

r+ \ / i

where p : ionic strength,N : integer depending on the electrolyte,

pK : equilibrium constant, andpK' : effective equilibrium constant in high p

solutions.

For H2CO3, N : 1 for pK1, and N = 2 for pKz.

As an example of the calculation, when p = 4.6the pK2 for carbonic acid is 8.95, compared to 10.33for very dilute solutions. This new point correspondsclosely with the beginning of poisoning of the calciterhombs. Hence, the abrupt decrease in solubilityof CaCOr results in more rapid crystallization andthe formation of the feathery and spicular,varietiesof calcite.

It should also be noted, however, that SiOe'2HzO(HaSiO+) becomes much more soluble in highlybasic solutions. For example, pK is 9.7 for the re-

action H+SiO+ € H* * HrSiO+, a point where thesolubility of SiOz increases appreciably. Thereforethe change in morphology of calcite may also berelated to the decreasing inertness (increasing solu-bility) of the silica gel.

Wood and Holliday (1960) discuss the insta-

bility of the HCOg- species in a heated aqueoussolution of carbonic acid open to the atmosphere:

2HCO.-(aq') : HzO * CO,t * CO,'-.

Since COz escapes from this system, essentially the

955

Frc. 8. Various morphologies of aragonite and vaterite:(A) pH : 7.3, (B) pH : 8.9, (C) pH : 8.9, (D) pH : 9.3,(E) pH : 9.3, (F) pH : 9.6, (G) pH : 9.6, (H) pH : 7.5,0.36 M NazCOs, 0.65 M CaCh (CC : calcite, A : aragonite'V : vaterite).

GROWTH OF CACO' PHASES BY THE SILICA GEL TECHNIQUE

flo o slr o,- .o

P H + -

Frc. 9. Slightly schematic solubility diagram

p H

for CaCO".

956 t. W. McCAULEY, AND R. ROY

same effect can be generated by raising the pH ofan HCO3 solution:

H C O 3 - : H * + C O r 2 - '

Hence, in uitro aragonite formation seems to dependon the presence of abnormally high concentrationsof COe'- with respect to the predominant HCO3-species.

Calcium bicarbonate [Ca(HCOg)2] exists in solu-tion, but no solid of this composition is known,although both CaCO3.H2O and vaterite may beara relation to this hypothetical material. Formation ofCaCOe in a bicarbonate solution involves the ejec-tion of a proton (H') from the growing crystal tomaintain charge balance in the solid:

Ca'* * Hcoa- : CaCosJ + H*.

Nancollas and Purdie (1964) point out that it isimportant to make allowance for these complexes(i.e., CaHCO3-) in crystal growth rate experiments;this was confirmed for CaCO3 precipitation byBischofi and Fyfe (1968). McCauley and Roy(1965) also suggested this as an important kineticconsideration. Hence, it is expected from purelykinetic arguments that appreciable amounts ofHCOB- will be incorporated into a rapidly enlarg-ing CaCO3 nucleus. The amount of entrapmentmust be dependent upon the nucleation rate and

the concentration of HCOg-. Nucleation rate is thatrate operative in a crystal nucleus before the criticalnucleus size is achieved and crystal growth takesover. Therefore, the preceding arguments indicatethat for certain pH regimes (or critical mole ratios ofCOr'- to HCO3-) the nucleation rate and HCO'-concentration are sufficiently high to result in ap-preciable entrapment of the latter into CaCO3. Thisentraprnent would seem to favor the substitution ofNa- for Ca2* to balance the excess positive chargecaused by the substitution of a HCO3- for a CO32-group.

Charge imbalance calculations (Pauling, 1929)would seem to indicate that appreciable substitu-tion of HCO.g- and Na* should favor a nine-foldcoordination of the cation by oxygen as in aragonite,thus leading to the preferential growth of aragonite-like nuclei rather than calcite. Hence, aragonite issuggested to be a function of proton entrapment,which can be controlled by the pH of the precipita-tion medium.

Direct evidence for the presence of protons inaragonite formed by this mechanism is hard toobtain. Buerger (1961) has indicated that wateranalyses on natural calcite and aragonite revealthat aragonite is always higher in water than iscalcite. Recently, Kinsman (1970) has shown thataragonite seems to incorporate abnormally high

3 4 5 6 7 8 I r O I 1 2 1 3m M o. + +m L o

( d l O 0 6 2 M M 9 C l 2 , O 3 5 M N o 2 C o 3 ,

p H : 7 8 5 , M E T H O D I I I

o 3 3

M o l e 7 o M g ( S r ) +

( o ) O 4 5 M C o C l 2 , 0 3 l M N o a c o 3 ,

p H = 7 7 5 , M E T H O D I

o 0 5 r . o 1 5 2 0 2 5 0 | 2

M o l e 7 " S r +

( b ) o 4 5 M C o C t 2 , o o z g M N o 2 c o 3 ,

p H = 8 2 0 , M E T H O D I

C A L C I T E

y' -s '

A R A G O N I T E

M o l € % M g ( S r ) +

( c ) o 0 9 M c o c t 2 , o o 7 8 M N o ? c 0 3 ,

p H : 7 7 5 , M E T I i O O I

C A L C I T E

A R A G O N I T E

I"" ' l

o/.'s olr

C A L C I T E / / -

, < s ,

Frc. 10. Percentage of calcite and aragonite as a function of Sr and Mg impurity concentration.

levels of Na*. In the present work, infrared absorp-tion spectra for aragonite formed by control of pHexhibit a weak band indicative of HCO3-. Finally,the proximity of aragonite formation to the car-bonate solutions further affirms the proposed mech-anism.

Influence of Impurity lons

Phase Distribution

The presence of strontium, magnesium, andnickel in the reaction medium favors the formationof aragonite, whereas barium impurities favor vater-ite. Figure 10 summarizes the results with strontiumand magnesium. It can be seen that the percentageof aragonite formed is directly dependent on theconcentration of the impurity ions and on the con-centration of the reactants. Note the dependence ofthe strontium effect on the relative concentrationof the reactants, aragonite formation being con-trolled by strontium more easily when the carbonateconcentration was much less than the calcium con-centration. Further, strontium always seems to bemore effective than magnesium in the formation ofaragonite. Figure 10d illustrates the effect of therelative concentration of calcium and magnesiumions on aragonite formation. In this series of runsthe concentration of Mg was held constant, whilethe concentration of calcium was varied, yieldinga total crystal yield which decreased appreciablywith decreasing Ca concentration.

The preceding results, however, must again beinterpreted in regard to Figure 6 which shows thathigh concentrations of reactants result in very largenumbers of nuclei and ensuing rapid growth; lowconcentrations of reactants have the opposite effect.For practically all of the impurity experiments, cal-cite formation (when it occurred) was primarilyrelegated to areas in the gel closer to Ca.

Figures 11A, B, and C are reflected light, planetransmitted, and crossed polarized light photomicro-graphs, respectively, of calcite crystals (distortedrhombs) and aragonite spherulites grown in thepresence of Mg. Calcite and' aragonite formed inthe runs summarized in Figure 10d are illustratedin 11D. In all of the impurity-generated aragonitespherulites, an optically distinct nucleus (core) wasdetected with the petrographic microscope. PowderX-ray difiraction analysis of hand picked spherulites(Table 2) revealed the presence of SrCO3 in Srinduced aragonite, but no identifiable phase within

957

w,Frc. 11. Influence of impurity ions on the morphology of

calcite and aragonite: (A,B,C) 8.92 m% Mg, (D) 0.16 MCa, 0.06 M Mg, (E) 2.5 mVo Sr, (F) 8.9 m% Ni, (G,H)

2.0 mVo Ni, ( I) 2.0 m% Ba. (CC - calcite, A = ara-gonite).

the Mg-aragonite. No foreign phase was found even

by X-ray analysis of Mg-aragonite spherulites with

their outer surfaces dissolved away.Several Sr-aragonite spherulites were cut in half

and examined with an electron microprobe. Figure

12 depicts Ca and Sr X-ray images of a "twinned"

spherulite showing SrCOa at the center of the

spherulite. An underexposed electron back scatter

Tlsrn 2. Percentage of Phases and Sr PartitionCoefficients

s r i n s r i n 5 r r n

n r . # * i e i Z e r a g . Z s r c o * * A r a g . D s . - A C C D s i c c

1 6 0 2 . 5 0 1 0 0 2 01 6 1 L , 2 5 r 0 0 1 51 6 2 0 . 6 3 9 0 1 52 8 3 0 . 2 5 8 5 1 0163 0" 13 852 8 4 0 . 0 5 1 5

* p H = 8 . 2 , 0 . 0 7 8 M N a 2 c o 3 , o . 3 8 M C a C f 2 ' r e t h o d I

**;st iated bg x-tal powdet di f f tact ion of ctushed aragonite

spheru- i i te***emission specttoscopi c anaTgses

GROWTH OF CACO: PHASES BY THE SILICA GEL TECHNIQUE

II.

A .

1 9 . 0 5 4 . 4 51 8 . 0 3 7 . 8 91 0 . 7 1 8 . 7 0 2 . 8 8 2 . 1 4

4 . 2 9 8 . 1 8 1 . 2 2 2 . 2 60 . 0 7 2 , 5 0 1 . 2 0 3 . 7 8

0 . 0 4 0 . 1 6

958 I. ll. McCAULEY, AND R. ROY

photograph (l2A) shows a small poisoned pris-matic crystal of SrCO3 in the center of the spherulite.

A group of hand picked Sr-aragonite spherulitesin various stages of development (Fig. 11E) illus-trate the proposed mechanism of formation. Earlystages of SrCO: formation and subsequent poisoningare labeled 1-4, while final development of variousshaped spherulites are numbered 5-7. This is sche-matically represented in Figure 13.

Similarly, Mg-aragonite spherulites were sec-tioned and analyzed with an electron microprobe.A Mg-rich phase was found in the center of thespherulites, but unequivocal identification was notpossible. Figure l4A, an electron back-scatter photo-graph of three joined spherulites, indicates the non-Ca phase at their centers. A Mg X-ray image of thismaterial (14B) showed this to be a Mg-rich phaseof very small size as compared to the relativelylarge SrCO3 nucleus. Exposing these small Mg-richnuclei for microprobe analysis proved to be ex-tremely difficult and was carried out successfullyon only two occasions. These data iUdicate that theinfluence of Mg is in terms of the formation of acertain number of nuclei and, rnt on the poisoningof calcite crystals, as suggested by Lippmann (1960)and reaffirmed by Bischoff (1968a).

The influence of Ni on the precipitation ofCaCOg is directly analogous to that of Mg, as ex-pected from crystal chemical considerations. Nickel,however, had adverse effects on the gel, limitingthis part of the study. Figure 11F is a reflected lightphotomicrograph of Ni-aragonite spherulites andcalcite crystals; the Ni-aragonite spherulites havecores similar to SrCOs (Fig. 11G). Incorporationof Ni into calcite crystals (Fig. 11H) produces apale green color.

Fto. 72. Electron probe X-ray and electron back scatterphotographs of a Sr induced aragonite spherulite: (A) elec-tron back scatter, (B) Ca X-ray scan, (C) Sr X-ray scan.Each division is 36 am.

! #_Jf_ +T i n y P r i s m o l i c P o i s o n i n g o n d s u b s e q u e n lS r C o 3 C r y s t o l r o d i o l q r o w t h o f o c i c u l o r c r y s l o t s

R o u n d i n g p r o c e s s o lo c i c u l o r c r y s f o l 3

F i n o l S p h e r u l i t e

Ftc. 13. Schematic representation of formation of aragonitespherulite.

Barium contaminants in the gel favor the forma-tion of isolated (non-overgrown) BaCO3 spherulitesand vaterite, and quite drastically alter the calcitemorphology; the latter effect can be seen in Figurelll As with nickel, however, barium producedadverse effects on the gel. In general, though, largeconcentrations of Ba in the reaction media ()0.1molal) resulted in the formation of calcite andwitherite (BaCO.g), while concentrations less thanthat resulted in no witherite and various amountsof calcite and vaterite. Apparently CaCO3 cannotovergrow on witherite for these conditions, as itcan on strontianite.

Hydrogen ion activity alters the effects of thevarious impurity cations on the formation of arago-nite and vaterite, but no definitive trends were noted.The precipitation of the various hydroxides is oneobvious consequence which must be reckoned with.Occasionally in high pH gels addition of the im-purity cations caused a milky-like appearance in thegel; this was thought to be either locally polymerizedgel or the hydroxide.

B. Partition of Impurily Cations Between GelSolution and Precipitate

Full understanding of the effect of impurity cationson the precipitation of CaCO3 in silica gel requiresa concurrent complementary study on the solid/liquid and solid/solid partition coefficients. Themagnitude of such an undertaking is considerableand, therefore, only preliminary results from amodest attempt are reported here. X-ray powderdiffraction and emission spectroscopic techniqueswere utilized to analyze for trace and minor im-purities in the solid phases. The different phasesand, in some cases, various morphologies of calcite

and aragonite were hand picked, repeatedly cleaned,and analyzed separately. Partition (distribution)coefficients (D) were calculated according to thehomogeneous distribution law as defined by Mclntire( 1 e 6 3 ) :where,

,-, _ Tr"/Cr"" -

Tr,fCr,

Tr" : trace component (moles) in solid,

Cr" : carrier component in solid :100/6-moLe /6 of traen component,

Tr1, Cr1 : similar concentrations in solution'

and,

D".-A : partition coefficient of Sr in aragonite,D'.-CC : partition coefhcient of Sr in calcite, andD*"-A, DM.-CC : identical values for Mg.

For this study D is employed as an efiectiuepartition coefficient, rather than an equilibrium one.Any comparison of these values to equilibrium onesis discouraged; rather, they should be utilized for

959

internal comparison only until they can be mean-ingfully normalized to equilibrium values. Further,all the analyses showed various amounts of sodiumand silicon, but these were neglected in the calcula-tions for the sake of simplicity.

Table 2 lists a representative sampling of parti-tion coefficients measured for Sr in aragonite andcalcite. Grand means for aII analyses (includingmany not listed in the table) showed the expectedresults:

D r " - A : 4 . 3 7 o : 2 . 3 5

D u . - C C : 1 . 0 4 4 : 1 . 0 8

Many factors collaborate to complicate the par-tition data. First, the partition of Sr into the varioussolids is a function of pH, since precipitation of thehydroxide will alter the epitaxy and incorporationmechanisms. For example, Ds"-A at a pH of 6.6is 25, whereas at pH 8.2 Ds"-A is 3. Second, variousmorphologies of calcite incorporate significantly

B

FIc. 14. Electron probe X-ray photographs of Mg inducedaragonite spherulite: (A) Ca X-ray scan, (B) Mg X-rayscan. Each division is 36 pm.

GROWTH OF CaCO' PHASES BY THE SILICA GEL TECHNIQUE

960 L W. McCAULEY. AND R. ROY

difierent amounts of Sr-averaging 1.7 mole percentSr in reaction area crystals and 0.9 in clear, un-distorted rhombohedra. Third, low cbncentrationsof reactants favorhigh Sr incorporations in aragonite,whereas high concentrations favor higli"Sr incorpora-tions in calcite. Finally, it must also be noted thatDs" is quite dependent on the relative percentageof simultaneously precipitating phases.

The data obtained on the effect of Sr on the pre-cipitation of CaCOg have reaffirmed the importantconcept that partition coefficients should reflect theamount of impuriiy cations in crystalline solution.Hence, the presence of SrCOs nuclei inclusionswill distort the resu{S if only bulk analyses arecarried out. Only procedures such as X-ray powderdiffraction, which measure the true amount in solu-tion, can measure equilibrium partition,cbefficientS.

In contrast to the inborporation of Sr iiito CaCO3,Mg showed no obvious overall preferenie for eithercalcite or aragonite. Further, some calcitb crystdlsgrowing in the presence of Mg had the beSt trans-parenqy of any calcite grown in this stlrdy. Tlieamount of Mg incorporated into calcite Should db-pend on the relative and absolute Mg arid Ca con-centrations. The data in Table 3, howevetu indicatethat Mg incorporation into calcite is mcjre depdndenton the relative percentages of calcite and aragonitenucleated. In other words, large amounts of Mgproduce Mg-rich nuclei upon which aragonite cangrow epitaxially. This, then, depletes the amountof Mg available for incorporation into calcite. Notethat Ca concentration was kept constant for theseruns. The data in Table 4 are for a series of runsfor which Mg concentration was held consiant,while Ca was varied. Both X-ray diffraction andemission spectroscopic analyses are listed for com-parison; the former being a more meaningful in-dication of true crystalline solution. Note also thatthe amount of Mg in calcite can be increased, for

TAsre 3. Percentage of Phases and Mg PartitionCoefficients

Tenrr 4. Various Analyses of Mg in CaCO,

Molezxx Mg MoIe z*x* MgRun /i* M C a M M C a . c c

3r03 1 13L23 1 3

0 . 3 0 0 . 2 00 . 1 6 0 . 3 70 . 0 6 1 . 0 00 . 0 3 2 . 0 0

3 0 0 . 3 5 4 . 9 2 . 63 0 5 . 53 0 0 . 4 1 5 . 3 1 , 55 0 0 . 8 6 1 1 . 3 6 , 5

nun,+* Ml.ln

z n.uo. Xnull orqo-n MBcin Duo-..

* p H . 7 . 8 , 0 . 0 6 M M q c 1 2 , 0 , 3 5 M N a 2 c o 3 , r e t h o d r l r**enjssion sprectto;copic analgses

* * * x-raa di f fraction ana-zqses

constant Mg concentratiolS, simply by decreasingthe Ca concentration. However, as soon as aragonitenucleation increases, less frlg will be availabie forincorporation into calcitO. The high Mg calcitesgrown in run #3lI are pictrtred in Figure 11D.All the preceding data strongly indicate that theeffect of Mg on the pfecipitation of CaCO3 is bythe nucleation of an Mg-rich phase upon whicharagonite can epitaxially grow, and not on thepoisoning of calcite nuclei.

C. Epitaxial Control ol Aragonite Formntion

Complementary X-ray diffraction and electronmicroprobe studies proved that Sragonite can growepitaxially on SrCO3 nuclei. MacDonald (1956)calculated that at least 30 riroli! percent of Sr inCaCO.r was needed to thermodynamically stabilizethe aragonite structure in the calcite sta'bility field.Ffowever, once aragonite (or an aragoniteJikephase) has reached critical nucleus (one which willpersist and grow) size, 30 mole percent Sr will nolonger be needed since aragonite can easily overgrowonto a Sr-rich ziragonite or Ca-rich SrCOs criticalnucleus. From a thermodynamic point of view,formation of the critical nucleus involves an in-crease in free eir.ergy, whereas crystal growth (after

nucleation) involves an exponential decrease in freeenergy and, thus, is a much more favorable thermo-dynamic step, allowing for the overgrowth of eithercalcite or aragonite, since there is negligible dif-ference energetically.

A similar mechanism must be operative in thecase of Mg, because of the similarity of the spheru-lites and the presence of :i Mg-rich phase in theircenter. Epitaxial control, requires at least two-di-mensional structural similarity. The crystal growth

of hexagonal KI on pseirdo-hexagonal mica is agood example of this. One measure of lattice param-

eter similarity is the percent misfit. Aragonite andSrCOs (strontianite) are quite similar (De Villiers,

r97 r ) :

M o l e % m o t e % * ' r i l o t e %

1 8 2 1 6 . 7] 8 5 9 . rt 8 6 4 . 82 9 2 I . 0293 0 .5

1 0 0 2 . 4 0 . 0 68 0 3 . 4 0 . ] 5 0 . 6 8 0 . 3 56 5 0 . 9 0 . 0 8 0 . 8 5 0 . 0 82 0 I . 4 0 . 6 5 1 . 0 0 0 . 4 52 0 ? . 7 2 . 5 3 2 . 6 1 2 . 5 0

* p H = 7 . 8 , 0 . 3 7 M C a C 1 2 , 0 . 3 1 M N a 2 C 0 3 , m e t h o d I* *em iss i on spec t roscop i c ana l yses

GROWTH OF CaCO' PHASES BY THE SILICA GEL TECHNIQUE 961

a(A) 6(A)strontianite 5.09 8.36aragonite 4.96 7 .97lo lNlisfrt 2.6 4.9

/oMisfit : (la, - a,l/a,) X 100,

c(A)6 .005 . 7 44 . 5

where, a, -- lattice parameter of nucleus.

Nesquehonite, MgCO3.3H2O, is the most likelyc,andidate as the Mg-rich phase. Davis (1906) wasthe first to point out that it instead of MgCO3precipitated from solution under normal conditions.Langmuir (1965) has recently confirmed these find-ings from precise thermodynamic data. The latticeparameters of nesquehonite (Stephan and MacGil-lavry, 1972) and aragonite agree quite well if the cparameter of aragonite is doubled:

a( ) afAl ce) pnesquehonite 5.37 7 .71 12.12 90"45'aragonite 4.96 7 .97 11 . 48 90o

/6 Misf i t 8 .3 3. 3 5.6

Therefore, it is proposed that nesquehonite is theMg-rich phase which nucleates first, allowing for thesubsequent epitaxial overgrowth of aragonite. Verysimilar mechanisms must be operative in the caseof Ni, but no nesquehonite-like Ni phase was found.

Applications to Geologic Conditions

The direct applications of these results to geo-logic environments is to be cautioned, althoughone can imagine the similarity of precipitation ina gel to precipitation in a thixotropic mud, in certaincave environments, in various types of sea animalswhere the gel mimics the organic part of the creature(e.9. echinoderm spines), in various sedimentaryrocks (limestone, shale, etc) where diffusion ofsolutions is slow, e/c. Nesquehonite, for example isundersaturated with respect to sea water, but localkinetic interlerences in equilibrium could affect theprecipitation of nesquehonite. Further, nesquehonitehas not been unambiguously identified in the arago-nite spherulites, thus other Mg-phases such as sepio-lite (which was critically examined) could be thekey phase. Nesquehonite seems to be the mostlogical, based on an examination on all knownMg-rich phases which could precipitate from theseenvironments. More detailed work is needed tosubstantiate this conclusion regarding nesquehonite.Follow-up work on the Ni systems might solve theproblem since Ni is much easier to detect with an

electron probe and the nuclei were bigger than forthe Mg-formed aragonites.

The formation of oolites, certain peculiar calcitemofphologies, and high Mg-calcites could be con-ceivably related to some of the observations madein this study. Thermodynamics is a very useful geo-logic tool, but many observed phenomena escapethermodynamic explanations and their interpreta-tion must be based on kinetic interferences inequilibrium processes. The writers strongly feel thatthe data presented in this paper might be veryrlsefully applied to certain unexplained geologicoccurrences of the various CaCOa phases, wherethermodynamics has failed. Further, the experi-lnental procedures established can allow for muchuseful follow-up work on CaCOa (especially onpartition of impurity ions) and other slightly solubleinorganic minerals. Other studies carried out onCaWOr and Nd2(COB)B'8H2O (McCauley andGehrhardt, 1970), CuzO-Cu (McCauley, Roy, andFreund, 1969), and unreported work on CaSOr'

2H2O (McCauley, unpublished) demonstrate the

flexibility of the experimental procedure and its

usefulness in studying the effects of precipitation

environment on the nucleation, crystal growth, and

doping of slightly soluble materials.

Acknowledgments

The authors would like to acknowledge the contributionof Professor Vladimir Vand (now deceased) to this work.He was responsible for re-introducing the very old gelcrystal growth technique as a viable mineralogical andrnaterials science tool to the authors and contributed greatlyin the preliminary stages of our learning process. The con-tributions of Dr. fames R. Fisher (USGS) in terms of ourunderstanding of aqueous geochemistry and CaCO, is alsogratefully acknowledged. The authors also wish to acknowl-edge the help of James R. Craig (VPISU) for his criticalreading of the manuscript. The study was carried out underARPA grant SD-132. The authors are also indebted to theArmy Materials and Mechanics Research Center for itssupport in the final stages of this work'

References

Brnre. C.. lNo J. Zeurrcre' (1971) Growth of CaCO' andCaSO*'2H"O crystals in gels. l. Crystal Growth, 10' 158-162.

Brsnorr, J. L. (1968a) Kinetics of calcite nucleation: Mag-nesium ion inhibition and ionic strength catalysis. .I.G e ophy s. Res. 7 3, 33 1 5-3322.

eNo W. S. FvrB (1968) Catalysis, inhibition andthe calcite-aragonite problem I. The aragonite-calcitetransformation. Am. l. Sci. 266' 65-79.

(1968b) Catalysis, inhibition and the calcite-

962 ,1. W. McCAULEY, AND R. ROY

aragonite problem II. The vaterite-aragonite transforma-tion. Am. L Sci. 266, 80-90.

BnIcxen, O. P., exo R. M. Gennpr-s (1967) Mineralogicfactors in natural water equilibria. In, Principles andApplications of llater Chemistry, S. D. Faust and J. V.Hunter, Eds., Iohn Wiley and Sons, Inc., New York,449-469.

Buencnn, M. J. (1961) Polymorphism and phase transfor-mations. Fortsch r. Mineral. 39, 9-24.

(1971) Crystal-structure aspects of phase transfor-mations. Trans. Am. Crystallogr. Assoc.7, i-23.

Drvrs, W. A. (1906) Studies of basic carbonates. l. Soc.Chem. Ind. 25,788-798.

Drrevstn, W. L., eNo L. Dscurrone (1950) Formation defa Calcite, Aragonite, et Vaterite. Bull. Soc. Chim. Belg.59. 40.

De VnuEns, J. P. R. (1971) Crystal structures of aragonite,strontianite, and witherite. Am. Mineral. 56.758-767.

DoNNAy, G., eNo J. D. H. DoNNey (1953) The Crystal-lography of Bastnaesite Parisite, Roentgenite and Syn-chisite. ln. Mineral. 38, 932-963.

FISHER, J. R. (1962) On the formation of calcite and ara-gonite: A review. The Netherworld News, 10, 103-114.

FrsnEn, L. W., exo F. L. SruoNs (1926a) Applications ofcolloid chemistry to mineralogy; part I. preliminary re-port. Ain. Mineral. ll, 124-130.

AND - (1926b) Applications of colloid chem-istry to mineralogy, Part IL Studies of crystal growth insilica gel. Am. Minerul. ll, 200-206.

Fnersrn, H., eNo Q. FrnNeNoo (1963) Ionic Equilibria inAnalytical Chemistry.. John Wiley and Sons, Inc., NewYork, 334 p.

Goto, M. (1961) Some mineralo-chemical problems con-cerning calcite and aragonite with special reference to thegenesis of aragonite. l. Fac. Sci. Hokkaido (Jniu., Ser.IV, 10, 571_640.

HeNrscu, H. K. (1970) Crystal Growth in Gels. ThePennsylvania State University press, University park,Pennsylvania, l l l p.

HorreNo, H. D., M. Boncsrr, J. MuNoz, AND V. M.Oxguncu (1963) The coprecipitation of Sr2* with ara-gonite and of Ca2* with strontianite between 90" andIOO'C. Geochim. Cosmrchim. Acta, 27,957-977,

H. J. HoruNo, exo J. L. MuNoz (1964) The co-precipitation of cations with CaCO3-II. The coprecipita_tion of Sr'?* with calcite between 90" and l}O.C. Geochim.Cosmochim. Acta, 28, 1287 -l3}l.

HoLMEs, H. N. (1917) The Formation of Crystals in Gels.J. Phys. Cltem. 21, 7O9-73j.

Huno, C. B., eNo R. L. Gnrrrrrn (1935) Studies on silicicacid gels V: The determination of the hydrogen ionconcentration of the gel mixtures. l. phys. Chem. 39,I 155-1 159.

IrBn, R. K. (1955) Tlrc Cotloid Chemistry of Silica andSilicates. Cornell University press, Ithica, New york,324 p.

JoxNsroN, J., H. E. MEnwru, eNo E. D. Wrr_lrlusow(1916) The several forms of calcium carbonate. Am. I.Sci. 41,473-512.

KrNsr"ren, D. J. J. (1970) Trace cations in araeonite. Geol.Soc. Am. Abstr. Programs,2, 596-597.

KrreNo, Y. (1962) The behavior of various ions in theseparation of calcidm carbonate from a bicarbonate solu-tion. Bull. Chem. Soc. Japan,,35, 1973-BA.

( 1964) On factors influencing the polymorphiccrystallization of calcium carbonate found in marinebiological systems. Recent Researchers in the Fields olHydrosphere, Atmosplrcre and Nuclear Geochemistry,Maruzen Co. Ltd., Tokyo, p. 305-319.

AND D. W. Hooo (1962) Calcium carbonatecrystal forms produced from sea water by inorganicprocesses. J. Oceanog. Soc., lapan,18, 141-145,

Konersu, I., rxo J. W. McCeumv (1973) Evidence ofcarbonate order-disorder in CaCO,,'H,O (abstr.). Am.Mineral. 5E, 1102

Ler*crrrurn, D. (1965) Stability of Carbonates in the SystemMgO-CO-HrO. lour. of Geology, 73, 730-754,

LIesEceNc, R. E. (1896) Chemishe Fernivirkurg. NochSep. von vf., Dusseldorf, Verlog von Ed. Liesegang. Phot,Archiv., 21,221.

LnrulNN, F. (1960) Versuche zur Aufklarung der Bil-dungs-Bedingunger von Calcit und Aragonit. Fortschr.Mineral. 38, 156-161.

MecDoNto, G. J. F. (1956) Experimental determinationof calcite-aragonite equilibrium relations at elevated tem-perature and pressures. Am. Mineral. 41r 744-756.

McCruI-Ey, J. W. (1965) Control of Nucleation, CrystalGrowth, and Doping of Various Calcium CarbonatePhases by the Gel Technique. M.S. Thesis, The Pennsyl-vania State University, University Park, Pennsylvania.

eNo R. Roy (1965) Gel-growth of pure and dopedCaCO" (calcite) and SrSOi single crystals (abstr.). Am.Ceram. Soc. Bull. (ab), 44,635-637.

AND - (1966a) Evidence for epitaxial controlof CaCO" phase formation as the mechanism of the in-fluence of impurity ions (abstr.). Trans. Am. Geophys.Union, 47, 2A2-203.

AND - (1966b) The effect of pH and concen-tration of reactants on the crystal growth of CaCO".Geol. Soc. Am. Spec. Pap. l0l, 736.

rNo H. Fneuxo (1969) Crystal growth ofcopper and nonstoichiometric cuprous oxide by the geltechnique. ACCG Conference on Crystal Growth, Na-tional Bureau of Standards, p. 40.

AND H. M. Gsnnnenor (1970) Crystal Growth olCaWO' and Nd, (CO,)a.\HO by the Gel Technique.Aurvrnc rn 70-13, Army Materials and Mechanics Re-search Center, Watertown, Mass.

McCoNNeu, J. D. C. (1960) Vaterite from Ballycraigy,Larne, Northern Ireland. Mineral Mag. 32, 535-544.

Mclxrrnp, W. L. (1963) Trace element partition coefficients-A review of theory and applications to geology.Geochim. Cosmochim. Acta, 27, l2A9-n64.

Meven, H. J. ( 1969) Struktur und Fehlordnung des Vaterits.Z. Kristallogr. lzt, 183-212.

Monse, H. W., AND J. D. H. DoNNAy (1931) Calcite arti-ficielle obtenue par diffusion dans un gel. BulI. Soc.Franc. Mineral 54, 19-23.

Munney, J. W. (1954) The deposition of calcite and ara-gonite in caves. ./. Geol. 621 481491.

Newcoues, G. H., exo N. Punore (1964) The kinetics ofcrystal growth. Chem. Soc. London Quart. Reu. lt, lJO.

Nrcrer, H. J., exo H. K. HrNrscs (1969) Growth of cal-cite crystals in gels. I. Electrochem. Soc. 116, 1258-1260.

Oxruncu, U. M., R. E. SrcNrr, eNo H. D. HoneNo (1959)Coprecipitation of strontium with calcium carbonate fromaqueous solutions (abstr.). BulI. Geol. Soc. Am. 70,1653-1654.

PAULING, L. (1929) The principles determining the struc-ture of complex ionic crystals. l. Am. Chem. Soc. 51,1010-1026.

Scnwenrz, A., D. Ecr,rnr, J. O'CoNNELL, AND K. FneNcrs(1971) Growth of vaterite and calcite crystals in gels.Mat . Res . Bu l l .6 , 134 l -1344.

Snlnur, B. D. (1965) Sodium bicarbonate and its hydrogenatom. Acta Crystallogr. 18, 818-819.

Srecrr, F. R. (1961) Variations of Sr/Ca ratios and Mgcontents in recent carbonate sediments of the northernFlorida Keys area. I. Sediment. Petrol. 31, 336.

Srvrrss, K. (1964) Variations in the crystalline form ofcalcium carbonate precipitated from artificial sea water.Nature. 201. 492-3.

963

Srenlt, F. G., eNo J. Howen (1961) Mineralogy and earlydiagenesis of carbonate sediments. l. Sediment. Petrol. 31,3 58 .

SrrrnlN, G. W., lNo C. H. MecGrluvnv (1972) Thecrystal structure of nesquehonite, MgCO"'3H"O. ActdCrystallogr. B2E, 1031-1033.

VeNo, V., H. K. HENrscH, AND J. W. McCeurev (1963)Growth of single crystals in gels (abstr.). Acta Crystal-logr. 16, A137.

Wooo, C. W., eNo A. K. Horuolv (1960) Inorganic Chem-isfry. Butterworths, London, 393 p.

Wn,c.v, J. L., lNo F. DeNrprs (1957) Precipitation of cal-ci te and aragonite. l . Am. Chem. Soc.79,2O3l-2O34.

Zzrrzn, E. J., eNo J. L. Wnev (1955) Factors influencingprecipitation of calcium carbonate. BulI. Am. Assoc.Petrol. Geol. 40, 140-152.

M.anuscript receioed, December 12, 1973; accepted

for publication, April 5, 1974.

GROWTH OF CaCO' PHASES BY THE SILICA GEL TECHNIOUE