American Mineralogist, Volume 72, pages842-844, 1987

Solubility and stability of scorodite, FeAsOo.2HrO: Discussion

Rorpnr G. RonrNs
Department of Mineral Processingand Extractive Metallurgy, University of New South Wales,
Kensington,N.S.W. 2033, Australia

Dove and Rimstidt (1985) reported solubility data for periments that take into account the srrggestedcom-
a synthetic scorodite prepared by mixing ferric chloride plexes.The stability region for ferric hydroxide, as shown
and sodium arsenate solutions and aging the resulting on this diagram, results from the use of a free energyfor
precipitatein the mother liquor at 95-105'C. The light FeOOH (amorphous)of -459.49 kJ'mol-', which would
tan-+olored product gave a distinct X-ray diffraction pat- be consistent with colloidal particles of about 5-nm di-
tern correspondingto scorodite. ameter. The incongruent point above which ferric arse-
The precipitated material was washed to remove elec- nate is unstable occurs here at pH : 4.8, which seemsto
trolytes and excessarsenateand was used in a seriesof be too high, as the divergenceofthe As and Fe concen-
solubility experimentsin which either HCI or NaOH was trations in a solution in equilibrium with ferric arsenate
used to adjust pH. The various mixtures were held at (Fig. l) indicates the incongruent point at pH : 2. This
25 + 0.5"C for periods up to eight weeksbefore analyzing point at pH : 2 could be obtained on the solubility dia-
the solutions for Fe and As and measuringpH. gram by assumingmore negativefree energyof formation
The actual solubility product and free energy of for- for amorphous FeOOH (i.e., larger particle size).
mation reported by Dove and Rimstidt are basedon the The solubility data obtained by Chukhlantsev(1956),
analysis of only four solution samples,all of which were Nishimura and Tozawa(1978),and Robins and Glastras
in the pH range 5.53-6.36. They reported a solubility (unpub.) are plotted on the solubility diagram in Figure
productK."(Fe',*AsOi-):l0-2' 7t05andaco : -1263.52 2. Dove and Rimstidt's results are also shown on the
kJ.mol '. diagram.
The solubility of ferric arsenate was determined by It is obvious that Dove and Rimstidt's results were
Chukhlantsev(1956) in solutionsthat were pH adjusted taken in a region where ferric arsenateexhibits incon-
with both sulfuric acid and nitric acid. Chukhlantsevana- gruent solubility, despitetheir claim to the contrary. The
lyzed the solutions only for Fe3+; the pH range of his scatter of their analyseson the four samplesis consider-
experimentswas 1.9-2.95,from which he reporteda sol- able and is possibly due to the fact that ferric arsenateis
ubility product K.. (Fe3*AsO; ) : l0-'z0'?41058.Dove and not stable in this region. The tan color of their product
Rimstidt claimed this solubility product to be too high, is also indicative ofincongruent origin, as ferric arsenate
probably owing to particle-sizeeffects. is off-white. However, their solubility product is fortu-
Nishimura and Tozawa (1978) reported a changein the itously closeto what might be consideredas a reasonable
solubility product of scoroditewith pH in the rangepH : value.Nishimura and Tozawa'ssingleexperimentat pH :
1-4, as is indicated in Table l. The calculated solubility 1.82grvesa solubility product of l0 2", which is more
product was based on the analysis of both Fe and As in reliably based as it seemsto have been obtained in the
solutions taken from both precipitation and dissolution region where ferric arsenateis stable. Chukhlantsev'sre-
experiments. A plot of the log of Fe and As concentra- sults in the region ofpH: 2 could also be treatedsepa-
tions vs. pH (takenfrom Nishimura and Tozawa)is shown rately to obtain a more reliable solubility product.
in Figure l; an obvious divergenceof the Fe and As levels
occurs at about pH : 2. The reason for this divergence
was later explainedby Robins (1981)as being due to the Tnele1. Solubility of FeAsOo.2HrO at 25"Cin dissolution and
instability of ferric arsenateabove pH : 2, where ferric precipitation (takenfromNishimura andTozawa,1978)
hydroxide becomesthe stable solid phase.The solubility FE As
of scorodite can therefore be regarded as incongruent Type of experiment Final pH (mg/L) (mg/L) -log K."
above pH : 2. Dissolution 1.82 T5 130 2 1. 1
Robins (1982) has shown the solubility of ferric arse- 2.06 20 91 21.3
nate (and other metal arseniccompounds) in the form of Fe:Asmole ratio 2.20 10 53 21.5
2.61 23 33 21.6
solubility diagrams (log activity vs. pH) and has shown
of FeAsO4: 1.02 288 0.41 29 21.8
(unpub. work) the effectof ferric arsenatecomplexessuch 310 0.17 25 22.2
as FeHAsO/ and FeH,AsO?* on the solubility of ferric Precipitation 2.03 41 79 21.0
arsenate.Figure 2 is a solubility diagram for ferric arse- 2.32 12 45 21.3
nate calculated by a computer mass-balancemodeling Initial[Fe] : 0.005 M 2.68 2.2 37 21.5
: 3.08 0.75
program and based on precipitation and dissolution ex- Initial[Asl 0.005 M 36 214

0003404x/87/0708-0842$02.00 842
 ROBINS: SCORODITE-DISCUSSION 843

+ CHUKIILAiSEY
Fo&4.
o q xlsHrllrR^rToz^wa
-5 \
o -u-r.---"r---r--""-- \ ^ A X DOVEIF|IST|DT
l! d-d \ (u
t-2 v \^ o O ROAIIISIGLASTBAS
o -a o
o -a E
E'\ogrc
o o -4
c\4.
O -s J
+Fc :@
J E
I
-5 filE E

-5 'EicO
-6

pH
Fig. 1. Fe and As concentrations associated with ferric ar-
senatein precipitationanddissolution;from NishimuraandTo-
zawa(1978) asin Tablel.
Robins and Glastras (unpub.) have obtained solubility
data for ferric arsenatein the pH range0-2,in the region
of congruent solubility. This work was done in perchlo-
rate, nitrate, and sulfate systemsand takes into account
complexing of Fe3* and As5*, as well as sulfate, nitrate,
and perchlorate complexes,to give a free energy of for-
mation for FeAsOo.2HrOof -1267.1 kJ'mol-'. Dove
and Rimstidt reported a free energy of formation for
FeAsOo.2HrOof -1263.52 + 0.28 kJ.mol-', assuming
no complexing of Fe3t and Ass*; however, such complex-
ing is considerable.I would suggestthat the free energies
of formations hown in Table 2 better explain the solu-
bility of scorodite(in perchloratemedia). Dove and Rim-
stidt suggestedthat the solubility and free energy of for-
mation given by them for ferric arsenateis more reliable
than the data that can be obtained from the work of
Chukhlantsev(1956).I would considerthe data from both
sourcesas inaccuratebecauseofhaving been determined
in the incongruent region.
Dove and Rimstidt also reported observations of As
mineral parageneses and statedthat there is a closespatial
TneLe2. Freeenergiesof formationfor the var-
iousspeciesconsideredin obtainingthe stability
diagramshown in Fig. 2
pH
Fig. 2. Computerstabilitydiagramfor ferric arsenate,cal-
culatedfrom the datain Table2 andshowingexperimental data
fromNishimuraandTozawa(1978),Chukhlantsev (1956),Rob-
ins andGlastras(unpub.),andDoveandRimstidt(1985).
relationship between arsenopyrite and scorodite. They
suggestedthat the reaction that forms scorodite is
FeAsS + l4Fe3+ + l0HrO -
l4Fe'z++ SO;- + FeAsOo'2HrO+ l6H*.
This equation may well represent a large part of an
overall oxidation mechanism,but the oxidation of arsen-
opyrite is far more complex and dependentvery closely
on pH. Other minerals formed as part of this mechanism
include orpiment and realgar(at low pH but in the region
where scoroditeis stable),but also pyrite, pynhotite, and
loellingite. The mechanism for the oxidation of arseno-
pyrite needs some attention, as does the mechanism of
its hydrothermalformation.
The stability of ferric arsenateis currently of particular
concern in relation to its disposal as a residuefrom min-
eral-extraction operations. In many locations, ferric ar-
senateis dumped into tailings and residue-storagesites,
on the assumption that it is stable and relatively insolu-
ble. The conversion of ferric arsenate(FeAsOo'2HrO) to
ferric oxyhydroxide (a-FeOOH) doesoccur however, with
the releaseof As to solution causinga potential environ-
mental problem.

No.* Species A4r*,u* (kJ mol-') Funrrmn DrscussloN

'I Fe3* - 16.99 The reply by Dove and Rimstidt (1987) to the above
FeOH'?* -238.49
-459 1I
discussionoftheir I 985 paper suggeststhat although their
3 Fe(OH),*
4 Fe(oH)8 -659.82 scorodite solubility measurementswere taken in the in-
5 Fe(oH); 830.06 congruent region where ferric hydroxide is the stable
o FeOOH(am) 459.49
Fe,(OHE+ -489 95 phase,the rate ofdecompositionofscoroditeis slow enough
I FeAsOo(S) -792.87 to allow meaningfulanalysisofboth As and Fe in solution.
o FeHAsO4+ -788.27 It has been my experiencethat if scorodite is placed into
10 FeHrAsOi+ -793.29
11 H3AsO! -766.01 a solution with a pH greater than 3, there is the almost
12 H"AsO; -753.16 immediate formation of a surfacelayer of what seemsto
13 HAsO?- -714.59
AsOl
be a less-densephasewith a distinctly reddish-brown col-
14 678.39
or. I have observedthesechangesusing transmissionelec-
. These numbers are used in Fio 2 to denote the
tron microscopy,and I believethe surfacelayer to be ferric
vanoussDecres.
hydroxide.
844 ROBINS: SCORODITE- DISCUSSION
The commentsby Nordstrom and Parks(1987)on the
Dove and Rimstidt (1985)papersuggestsomeaspectson
which I can elaboratefurther.
Nordstrom and Parks(l 987) agreethat the regionwhere
scorodite will exhibit incongruent dissolution is where
ferric hydroxide becomesthe stablephase,and that region
can be delineatedwith a pK,. for "ferric hydroxide." Lit-
erature values for pK"", accordingto Flynn (1984), are in
the range 37.1-39.0 for fresh to agedprecipitate and, ac-
cordingto Langmuir and Whittemore (1971),dependon
particle size. The literature values cited were originally
obtained for precipitation of ferric hydroxide from pure
solutions, not for a surface decomposition of one crys-
talline material (scorodite) to yield ferric hydroxide.
Nordstrom and Parks (1987) also suggestthat the decom-
position of scorodite may occur with the formation of a
surfacecoating of ferric hydroxide (I find this to be the
case,as previously stated). Under these conditions, the
apparent pK". for ferric hydroxide is likely to be quite
different from the values for a precipitate formed directly
from solution, perhaps influenced by the particle size of
the underlyingscorodite.
In my laboratory, my coworkersand I have studied the
formation and decomposition of scorodite fairly exten-
sively. We find that when scorodite is precipitated at a
pH just below the "incongruentpoint" (pH e 1.8),it has
a negative surface charge. If the pH is then raised, the
surfacepotential becomespositive, and an examination
of the precipitate by transmission electron microscopy
(using a technique similar to that describedby Green et
al., 1974) would indicate that at this stage,the surfaceof
the scorodite is coatedwith a separatephase,assumedto
be ferric hydroxide. A separateinvestigation that involves
the use of attenuated total-reflection Fourier-transform
infrared spectroscopyalso indicates the formation ofsur-
face ferric hydroxide. The comments by Nordstrom and
Parks are certainly speculative,but I believe accurate.
We have also investigated the precipitation of ferric
hydroxide onto scoroditeparticlesaswell as onto a variety
ofother crystallinematerials.We find that the pH at which
surfaceprecipitation of ferric hydroxide occurs depends
on the surfacepotential and particle sizeofthe other phase
and that in each case the ferric hydroxide can be char-
acteized by a different "pK"." value. In the caseof ferric
hydroxide on scorodite,"p(o" : 39.5.
Nordstrom and Parks also comment on the use of free-
energy values versus pK,. values. This really is a matter
ofconsistency ofdata, and provided that the free energies
ofall ofthe speciesin the given systemare quoted,then
it is perhaps easier to manipulate free-energydata that
relate to the individual componentsof the system.Many
pK"" data have beencalculatedfrom free energies,as were
thosefrom Flynn (1984) quoted above. Dove and Rim-
stidt (1985)did not give all free energiesin the systemin
specifying a free energy of formation for scorodite. My
own preferencefor free energiesis becausethe computer
programs that we have developed for manipulating the
data and plotting stability diagrams require an input of
free energiesof formation for the various speciesin the
system.There is also, of course,the problem of correction
for ionic strength (or correction to the molal infinite di-
lution standardstate)which, in a systemsuch as the Fe3t-
As5+-HzOsystem with the presenceof various complex
"ion pairs," becomesdifficult.
RsrnnnNcns
Chukhlantsev, V.G (1956) Solubility products of a seriesof arsenates.
Zhurnal AnalitichesikoiKhimii. I l. 565-571.
Dove, PM., and Rimstidt, J.D (1985) The solubility and stability of
scorodite,FeAsOo 2H,O. American Mineralogist, 70, 838-844.
-(1987) The solubility and stability of scorodite,FeAsOn2HrO.
American Mineralogist, 72, 845-848.
Flynn, C M (1984)Hydrolysisofinorganic iron(Ill) salts.ChemicalRe-
views,84,3l-41
Greene,R.S.B.,Murphy, P.J , Posner,A.M., and Quirk, J.P. (1974) A
preparativetechnique for electron microscopic examination of colloid
particles.Claysand Clay Minerals,22, 185-188.
Langmuir, D., and tvvhittemore,D.O. (1971) Variations in the stability
ofprecipitated ferric oxyhydroxides.In J D. Hem, Ed., Nonequilibrium
systemsin natural water chemistry, p. 209-234 American Chemical
Society,Advancesin Chemistry Series106, Washington, D C
Nishimura, T., and Tozawa,K ( 1978) On the solubility productsof ferric,
calcium and magnesiumarsenatesBulletin ofthe ResearchInstitute of
Mineral Dressing and Metallurgy (Tohoku University), 34,20-26 (in
Japanese).
Nordstrom,D.K., and Parks,G A (1987)The solubilityand stabilityof
scorodite,FeAsOn2H,O: Discussion.American Mineralogisl,,72, 849-
851
Robins, R.G. (1981) The solubility of metal arsenates.Metallurgical
TransactionsB. l28, 103-109.
-(1982) The stabilities of arsenic(V)and arsenic(Ilf compounds in
aqueousmetal extraction systems.In K. Osseo-Asareand J.D. Miller,
Eds., Hydrometallurgy research,development and plant practice, p.
291-310. The Metallurgical Society of AIME, Warrendale, Pennsyl-
vanla.
MaNuscmpr RECETvED JeNulnv 28, 1986
Meuuscnrpr AccEprEDJnNuenv 26, 1987
Funtrrn DrscussroNRECEIVEDMev 5, 1987