I R R E V E R S I B I L I T Y O F S U L F A T E S O R P T I O N O N

G O E T H I T E AND H E M A T I T E

LAURIE J. TURNER* and JAMES R. KRAMER

Department of Geology, McMaster University, 1280 Main Street West, Hamilton, Ontario, Canada L8S 4M1

(Received January 28, 199i; revised July 2, 199I)

Abstract. Sulfate ion adsorption and desorption experiments carried out on synthetic goethite and hematite and natural hematite show adsorption to be a highly irreversible reaction. All oxides showed an increase in sulfate ion adsorption with decrease in pH. Only a small fraction of sorbed sulfate was desorbable after 48 hr, and only at a pH of 3. Extreme irreversibility of sulfate sorption on these common soil minerals suggests that adsorbed sulfate is more immobile in watersheds than previously considered and that recovery models which inherently assume reversibility may need to be modified.

1. Introduction

The Direct Delayed Response Program (DDRP) has been employed to assess the

long term effect of acid deposition. Indeed much of the Acid Rain literature

conclusions are geared towards a saturation-desorption phenomenon. Inherent in

the models (i.e. MAGIC) is the assumption of reversibility of sulfate adsorption

(Cosby et al., 1986; Hornberger et al., 1986), even though some literature suggests

that sulfate sorption is more irreversible in nature.

Sulfate ion retention in a soil depends on environmental conditions and site reac-

tion characteristics, with oxyhydroxide and clay minerals being important soil con-

stituents. Definition of constituent sorption charateristics and consideration of factors

which influence sorption are prerequisite to the modeling of sulfate ion sorption

in soils. When the sorption process on individual soil constituents is ascertained,

then parallel effects in whole soil behavior may be estimated. Natural whole soil samples are a complex mixture of surfaces with a variety

of organic and inorganic species in the soil solution. Many soils have been examined for sorption characteristics (Kamprath et al., 1956; Liu and Thomas, 1961; Chao

et al., 1962; Chang and Thomas, 1963; Barrow, 1967; Bornemisza and Llanos,

1967; Hasan et al., 1970; Sanders and Tinker, 1975; Couto et al., 1979; Singh,

1984), but derivations of mechanisms from soil studies is difficult due to their complex

nature, and the inability to specify specific reactions with certainty. The sorption

mechanisms of the individual constituents can be interpreted with less ambiguity,

since specific mineral components and surfaces can be more thoroughly characterized.

Extrapolation of experimental results for individual surfaces must be done with

caution, however, because when placed back into the natural environment, con-

* Current address: 10 York Street, Dundas, Ontario, Canada L9H 1L2.

Water, Air, and Soil Pollution 63: 23-32, 1992. �9 1992 Kluwer Academic Publishers. Printed in the Netherlands.

24 LAURIE J. TURNER AND JAMES R. KRAMER

stituent behavior may be modified by interaction with other substrate components. Anion sorption on oxyhydroxide minerals has been studied fairly extensively

(Bowden el al., 1973; Breuwsma and Lyklema, 1973; Parfitt and Russell, 1977; Parfitt and Smart, 1977, 1978; Sigg and Stumm, 1981; Borggaard, 1983), but only a few (Aylmore et al., 1967; Hingston et al., 1974; Russell et al., 1974) have examined desorption also. This paper present results of sorption-desorption studies on the iron oxide, hematite, and hydroxide, goethite. These important soil minerals are

easily prepared and known to sorb sulfate.

2. Experimental Methods

P R E P A R A T I O N OF MATERIALS

Goethite was prepared by aging a solution of ferric nitrate, adjusted to pH 12, at 60 ~ for 24 hr (Atkinson et al., 1968). Hematite was prepared by hydrolyzing ferric nitrate at 100 ~ under reflux conditions for 18 d (Parks and DeBruyn, 1962). Precipitates were washed, filtered and vacuum dried. X-ray diffraction analysis (XRD) indicated that the materials obtained were goethite in the first case, and hematite with a small amount of goethite impurity in the second case. A sample of natural hematite was crushed and sieved. XRD analysis showed no crystalline impurities.

The surface areas of synthetic goethite and hematite, measured using the BET- N2 adsorption method (Brunauer et al., 1938), were 44.1 and 43.2 m 2 g l, respectively,

while the crushed natural hematite sample was 12.0 m 2 g 1. The pH of zero points of charge as determined by acid/base titrations were

7.1, 8.0, and 8.8 (all + 0.05) for goethite, hematite and natural hematite, respectively. All chemicals used were of analytical reagent grade.

SORPTION EXPERIMENTS

Ooethite and hematite samples (0.1 g) were mixed with 25 to 50 mL of K2NO3 media of 10 -2 to 10 -4 M concentration. These constant ionic strength suspensions were equilibrated for 24 hr at a constant pH value. Known amounts of sulfate were added K2SO 4 to attain initial solution concentrations up to 5 mM. At pre- set time intervals (0.5 to 48 hr), the pH was recorded, and a sample of the suspension removed and filtered through 0.45 v.m Millipore filter paper. The filtrate was analyzed for sulfate using ion chromatography.

D E S O R P T I O N EXPERIMENTS

Reversibility was examined by leaching in a constant pH and ionic media. This was achieved by filtration and resuspension of a sample to which a known amount of sulfate had been sorbed. The ionic media and pH were the same as for uptake studies, and the ratio of solid to solution was kept constant. After a reaction time of 48 hr, the suspension was filtered and the filtrate examined for sulfate con-

IRREVERSIBILITY OF SULFATE SORPTION 25

i f )

i

O E

O Or)

O E =,,

11)

i f ) Q)

i

i f )

O E =k

300

250

200

150

100

A. Goethite

t ~ 0 0

rn 17

s 0 , ~

~ ~ 4 ' ~ 6 Equilibrium Solution Concentration (retool/I)

300

250

200

150

100

B. Hematite.?J ................

s~- , , , , ~ 6

Equilibrium Solution Concentration (mmol/1)

300 C. Natural Hematite

250

200

150

100

50.

0

E(

p~3

~'~ ~, ~ 6 uilibrium Solution Concentration (mmol/i)

p~5 pl~17 Des.9~, l!on

Fig. 1. Adsorption isotherms for sulfate sorption on (A) goethite, (B) hematite and (C) natural hematite at room temperature for pH 3, 5 and 7. Desorption data are indicated by an asterisk.

26 LAURIE J . T U R N E R A N D JAMES R. KRAMER

centra t ion. The entire p rocedure was repea ted several t imes unti l no add i t iona l

sulfate could be detected in solut ion. The var ious SO4 fract ions were summed to

de te rmine the to ta l desorbab le fract ion.

3. Results a n d D i s c u s s i o n

SULFATE SORPTION

The so rp t ion process was in i t ia ted by a r ap id reac t ion (~1/2 hr) fo l lowed by a

longer slow one which reached a s teady state or a p p a r e n t equi l ib r ium in 24 hr.

3

2.5

E 2

"~1.5 0

"J 1

0.5

0

A Hematite . . . . . . . . . . s . . . . . . .

-1~ . . . . . . . . . . . . . . . . . . ~ ' " ............... ~.:~:.-: ................................. = .~ .~ . .~~. - . . . . - . �9 .:.~77=~" * ' ' . . . . . . . . .. 0 . - . . - - - - ~ ' ' U "

I I I 1 I 1.5 2 2.5 3 3.5 4

Log Eqilibrium Sulfate Concentration, pmol/I

2 . 5

B Goethite D .... . . . . . . . . . .E3--'s ......

2 . . . . . . . . . . . . ~ - . . . . . E} . . . . . . . . . . . .

E . . . . . . --.T

o, ~ " " ' " ~ o . . fo

0.5 t i t i i 1 1.5 2 2.5 3 3.5 4

Log Eqilibrium Sulfate Concentration, )Jmol/I ..~.. z~_ _~_ ~"~

Fig. 2. Freundlich plots for sulfate sorption on (A) hematite and (B) goethite at three pH levels. NH = Natural hematite.

IRREVERSIBILITY OF SULFATE SORPTION 27

That is, no further changes in pH or sulfate concentration were observed after

24 hr. This rapid equilibrium attainment is a c o m m o n phenomenon of anion sorption experiments. In all cases, the oxides sorbed increasing amounts of sulfate as solution pH was decreased. The amount sorbed also increased with solution sulfate con- centration. There was no effect of changing ionic strength on adsorption for hematite, but goethite sorption increased with increasing ionic strength at a pH of 5. This is in contrast to the observations of Barrow (1967) who noted a decrease in sorption with increasing ionic strength up to 0.01 M, and no effect above this concentration. No solid concentration effect was observed in any of the studies.

Figure 1 shows the sulfate adsorption isotherms obtained for the three samples

120

100

8O

E 6o

o

A Hematite / / . t . . . .

j .," �9 ,,"

~ ,. '" / ....*'"

J ,,.,,., .. ,o�9149149 ,~ S ~ /* / / ~ ~ ~,, "~ S j"

40 , / . ...... . . . . . . ~ / . / .- .... . . - "

2 0 ' . . . . . . . . . . . . 1~. . . . . . . . . . . . . . . . . . .

o . . . . : . . . . . . . . . . . . . . . . i , , 0 1,000 2,000 3,000 4,000 5,000 6,000

Eqilibrium Sulfate Concentration, ,umol/i

80

6O

E 40

o

/ /

1

.s J

t

B Goethit~ /

/ /

/ i

/

r ,5-"" / / ~ o O ~ 1 7 6 1 7 6

/ 4" o 1 ~ "~ �9 , j o . o ~

2 ..,, 1~ . . . . . . .

I I I 1 0 1,000 2,000 3,000 4,000 5,000 6,000

Eqilibrium Sulfate Concentration, jJmol/I

Fig. 3. L a n g m u i r plots for sulfate sorpt ion on (A) hemat i te and (B) goethi te at three p H levels. N H = Natura l hemati te .

28 LAURIE J , T U R N E R AND JAMES R. KRAMER

under s tudy. The p r e p a r e d goethi te reached an appa ren t so rp t ion p la teau at all

p H levels. The na tu ra l hemat i te sample also reached a so rp t ion max imum. The

synthet ic hemat i te , however , a p p r o a c h e d a so rp t ion m a x i m u m only at p H 7. The

lower p H exper iments tend to exhibit a cont inuing increase in sulfate with sulfate

ion concen t ra t ion .

So rp t ion da ta for 24 hr exper iments were c o m p a r e d to bo th the F reund l i ch and

L a n g m u i r type equat ions (Figures 2 and 3). F i t t ed pa rame te r s and the goodness-

of-fit for specific p H s are summar i zed in Table I.

Coefficients of va r ia t ion indicate that for p H 3 and 5 the L a n g m u i r form of

equa t ion is a bet ter fit to the exper imenta l da t a than the F reund l i ch equat ion.

At p H 7, however , coefficients of va r ia t ion favor the F reund l i ch equa t ion even

though the cor re la t ion coefficients are poo re r than for the L a n g m u i r fits. Overal l ,

a d s o rp t i on i so therms for the oxides examined fol low the L a n g m u i r equa t ion

TABLE I

(a) Langmuir equation parameters for Oxide Samples, X / M - ( b C ) / ( K + C ) . The inverse linear form was used for fitting, and the coefficient of variation (C.V.) is determined by (X/ (E(I ' - I~m)2/n) ) / I~ m where 12, Fro, and n are predicted uptake, mean difference between predicted uptake and experimental uptake

and number of samples, respectively. Sorption units are in txmol L -1 sulfate ion and ismol g-1 of solid

Oxide Correlation Coefficient Sorption Equilibrium sample pH coefficient of variation maximum constant

Goethite 3 0.9999 0.5010 152.3 32.9 5 0.9992 0.5729 74.6 186.5 7 0.9980 0.5642 23.7 130.5

Hematite 3 0.9828 0.3625 278.4 408.3 5 0.9750 0.5430 82.4 603.6 7 0.9977 0.5799 20.2 127.6

Natural 3 0.9990 0.2931 45.9 173.1 hematite

(b) Freundlich equation parameters. Parameters are obtained from the linear logarithmic transformation of the equation, X / M = k O/P -. X / M is in ~mol g-~ and C in ~smol L -~

Oxide Correlation Coefficient sample pH coefficient of variation p k

Goethite 3 0.9504 0.8357 0.1047 1.8 5 0.9567 0.5915 0.2160 1.1 7 0.9858 0.3465 0.4229 0.1

Hematite 3 0.9414 0.4435 0.1961 1.7 5 0.9433 0.5966 0.2237 0.8 7 0.9563 0.2294 0.4729 0.1

Natural 3 0.9794 0.9346 0.1222 1.2 hematite

IRREVERSIBILITY OF SULFATE SORPTION 29

reasonably well over all concentration ranges studied,. The sorption maxima for goethite listed in Table I agree fairly well with the observed maxima of 151.0, 72.9, and 20.8 p.mol g 1 for pH levels 3, 5 and 7, respectively. No maxima were observed at pH 3 or 5 for the synthetic hematite for the range of concentrations under study, however, a maximum of 17.6 tsmol g< observed at pH 7 is within range of the calculated Langmuir maximum. For the natural hematite sample, the calculated adsorption maximum is in fair agreement with the observed maximum of 45 gmol g 1.

Other studies also indicate a Langmuir form of equation can be fitted to sulfate adsorption data, because a plateau or maximum is achieved. Hingston et al. (1972) reported a sorption maximum of 150 p.mol g-~ at pH 3 for a prepared goethite sample. Parfitt and Smart (1977) determined the sorption maximum for a prepared goethite to be 75 p.mol g-1 and 125 ~xmol g-I for pH 5.1 and 3.4, respectively. A maximum of 67 ~xmol g< was reported by Aylmore et al. (1967) and 85 p.mol g t by Parf t t and Smart (1978) for a prepared hematite at pH 4.6 and 3.5, respectively. Wootton (1985) determined the apparent sorption maximum for the same natural hematite sample used in this study to be 46 ~mol g ~ at pH 3.0.

The values for sorption maximum reported in the literature and here are normalized to BET-N2 specifc surface areas in Table II. Differences in maximum surface saturation are directly attributable to variations in measured or calculated specific surface area. Goethite samples in this study exhibit half and hematite samples about twice the specific surface area of samples reported in the literature. Aylmore et

a/. (1967), Hingston el al. (1972) and Wooton (1985) used a BET-N2 method for measuring specific surface area. Parfitt and Smart (1977, 1978) calculated the surface area for their goethite and hematite particles from phosphate adsorption data.

TABLE II

Experimental and reported sorption max imum for oxide samples normalized to specific surface areas

Maximum Area Maximum Sample pH fsmol g-~ m 2 g ~ p~mol m 2 Reference

Goethite 3 151.0 44.1 3.42 this study 3 150 81 1.85 Hingston et al., 1972 3.4 125 90 1.39 Parfitt and Smart, 1977 5 72.9 44.1 1.65 this study 5 60 81 0.74 Hingston et al., 1972 5.1 75 90 0.83 Parfitt and Smart, 1977 7 20.8 44.1 0.47 this study

Hematite 3.5 85 22 3.86 Parfitt and Smart, 1978 4.6 67 26.7 2.50 Aylmore et al., 1967 7 17.6 43.2 0.41 this study

Natural 3 45.0 12.0 3.72 this study hematite 3 46.0 12.0 3.80 Wooton, 1985

30 LAURIE J. TURNER AND JAMES R. KRAMER

SULFATE DESORPTION

Sulfate sorption is strongly pH dependent and largely irreversible (Figure 4). The

sorption pH also appears to control the desorption. Sulfate sorbed at low pH was

generally more desorbable than that at higher pH (0%) due to being less strongly

bound. In fact, only those sorption experiments at pH 3 showed any measurable

desorption. Only 41, 29, and 6.5% of sorbed sulfate is desorbable at pH 3 from

goethite, hematite, and natural hematite, respectively.

(D .Q 50 L _

o A if}

40 a

10

0 2

3o CO o 20 d~

c-

O

13_

-o 70

c, 60 I1,=,

O 50 a . 40

3 0 = ," O)

20

o 10 - E =1-" 0

2

Goethite

.!1 o~176176

~176 ~176176176176 o.~ ~

,~176176176 o~176176 ~.~176176176

O ~176176176176176176176176176176176176176176176 ...... ~176176176176176176176176176176176176

3 4 5 6 7 pH of Desorption

8

B . . . l ~176

~176 o."~,.~ ...:/""

..'S"" .o ~

o.'..o

....222"" .,..222""" ~::..'"

t - . . ..... + ..... :'"'; ..... :'"'T'": "''~

3 4 5 6 7 pH of Desorption

He.~tite Natural Hematite . . . . . . . 0 . .

8

Fig. 4. pH dependence of sulfate desorption from goethite, hematite, and natural hematite. The sulfate was sorbed at pH 3 in order to saturate the surface before desorption. (A) pH vs pecentage of sulfate

desorbed. (B) pH vs p, mol g i of sulfate desorbed.

IRREVERSIBIL1TY OF S U L F A T E SORPTION 31

Increasing amounts of sulfate were released as desorption pH increased. Figure

4a shows that goethite desorbs the largest percentage of sorbed sulfate, and natural

hematite the smallest. Figure 4b indicates that goethite and hematite actually desorbed about the same absolute amount of sulfate (60 txmol g-X), whereas the natural hematite sample released much less (4 gmol g-l).

4. Implications For Modeling

In modeling watershed recovery processes, the DDRP describes the dynamic response of surface water chemistry as a function of rates of acidic deposition and several key soil processes. In particular, soil sulfate adsorption is used as the controlling factor of long term sulfate dynamics (Cosby et al., 1986; Hornberger et al., 1986). Inherent in the model is the assumption of reversibility of sulfate adsorption;

equilibrium adsorption and desorption. The present work has shown sulfate adsorption on iron oxides to be for the most part irreversible, with desorption only occurring under specific conditions. It is important to realize that iron oxides are not soils, but important soil minerals. In addition to iron oxides, other minerals (such as aluminum oxides) play an integral part as substrates for the adsorption of sulfate. If aluminum oxides behave in the same manner as iron oxides, we hypothesize that soil may behave as follows:

A soil with a measurable amount of sulfate sorption capacity will follow the DDRP model when amounts of acidic deposition are increased. That is, incoming sulfate will adsorb onto soil oxides delaying the equilibration of surface water chemistry until the sorption capacity is reached, after which hydrologic retention

will determine the remaining time to equilibrium. However, because of the irre- versibility of sulfate sorption, the system will not react as predicted when acidic

deposition is decreased. A neutral or alkaline soil is not likely to adsorb a large amount of sulfate. An

acidic soil, on the other hand, will adsorb a larger amount if sulfate and has the best chance of desorbing some of that sulfate. Present data indicate that the amount

desorbed will be small if the pH remains low and will only increase if the pH is raised. If a small amount of sulfate is desorbed, there will be a time lag greater than that of hydrologic response time before equilibrium is again reached. However, if no sulfate is desorbed, as in the case of neutral and alkaline soils, equilibrium will be established according to hydrologic response time alone.

5. Summary

The sorption and desorption of sulfate ions on hematite and goethite have been examined at equilibrium concentrations up to 5 mM SO4 =. Both sorption and desorption are highly pH dependent. Sorption increases with decreasing pH, as seen in previous studies with iron oxyhydroxides and soils, with the maximum amount adsorbed (saturation) occurring near a pH of 3. Previous studies with soils

indicate Langmuir behavior (with its sorption maximum) in some cases, and

32 LAURIE J. TURNER AND JAMES R. KRAMER

Freundlich behavior (with no sorption maximum present) in others. Present sorption isotherms are in better agreement with Langmuir behavior. In this study, a sorption maximum or plateau is reached at all pH levels examined except for synthetic hematite which reached maximum at pH 7 only. The sorption maximums measured here for goethite agree particularly well with maxima from other studies of this mineral.

Soil studies indicate sorbed sulfate to be partially desorbable, with amounts desorbed dependent upon soil properties such a mineral content. In the present study, Aylmore et al. (1967) initial indication of irreversibility of sulfate sorption on iron oxyhydroxide minerals is confirmed. Under present conditions, only sulfate sorbed at pH 3 is labile, with amounts desorbed increasing with the pH of the desorption solution. A maximum of 41, 29, and 6.5% of total sorbed sulfate can be desorbed from goethite, hematite and natural hematite, respectively. This lack of reversibility has important implications for models which simulate watershed response to changing inputs of acidic sulfate deposition. Sorption onto soil minerals is assumed to be reversible, and watershed recovery models were formulated around this assumption. In fact, sorption onto iron oxide minerals is not reversible, and present models must be modified to reflect this irreversibility. All substrates (including whole soil samples) need to be investigated before assumptions about soil behavior can be assessed quantitatively.

References

Atkinson, R. J., Posner, A. M., and Quirk, J. R: 1968, J. Inorg. Nucl. Chem. 30, 2371. Aylmore, L. A., Karim, G. M., and Quirk, J. R: 1967, SoilSci. 103, I0. Barrow, N. J.: 1967, Soil Sci. 104,342. Borggaard, O. K.: 1983, Clays Clay Mins. 31,230. Bornemisza, E. and Llanos, R.: I967, Soil Sci. Soc. Am. Proc. 31,356. Bowden, J. W., Bolland, M. D. A., Posner, A. M., and Quirk, J. P.: 1973, Nature Phys. Sci. (London)

245, 81. Breeuwsma, A. and Lyklema, J.: 1973, J. Coll. Int. Sci. 43, 437. Brunauer, S., Emmett, R H., and Teller, E.: 1938, a'i Am. Chem. Soc. 60, 309. Chang, M. L. and Thomas, G. W.: 1963, Soilgci. Soc. Am. Proc. 27, 281. Chao, T. T., Harward, M. E., and Fang, S. C.: 1962, Soil Sci. Soc. Am. Proc. 26,234, Cosby, B. J., Hornberger, G. M., Wright, R. F., and Galloway, J. N.: 1986, Water Resour. Res. 22,

1283. Couto, W., Lathwell, D. J., and Bouldin, D. R.: 1979, Soil Sci. 127, 108. Hasan, S. M., Fox. R. L., and Boyd, C. C.: 1970, SoilSci. Soc. Am. Proc. 34, 897. Hingston, F. J., Posner, A. M., and Quirk, J. R: 1972, J. Soil Sci. 25, 16. Hingston, F. J., Posner, A. M., and Quirk, J. E: 1974, J. Soil Sci. 25, 16. Hornberger, G. M., Cosby, B. J., and Galloway, J. N.: 1986, Water Resour. Res. 22, 1293. Kalnprath, E. L., Nelson, W. L., and Fitts, J. W.: 1956, Soil Sci. Soe. Am. Proc. 20,463. Liu, M. and Thomas, G, W.: 1961, Nature 192, 384. Parfitt, R. L. and Russell, J. D.: 1977, J. SoilSci. 28, 297. Parfitt, R. L. and Smart, R. St. C.: 1977, J. Chem. Soc., Far. Trans. 73, 796. Parfitt, R. L. and Smart, R. St. C.: 1978, SoilSci. Soc. Am. J. 42, 48. Parks, G. A. and DeBruyn, P. L.: 1962, J. Phys. Chem. 66,967. Russell, J. D., Parfitt, R L., Fraser, A. R., and Farmer, V. C.: 1974, Nature 248, 220. Sanders, F. E. and Tinker, R B. H.: 1975, Geoderma 13, 317. Sigg, L. and Stumm, W.: 1981, Coll. Surf 2, 101. Singh, B. R.: 1984, SoilSci. 138, 189. Wootton, R. S.: 1985, BSc. Thesis, McMaster University, Hamilton, Ontario.