Applki Geochemisfr.v,Vol. 10. pp. 715-723.

1995
Pergamon Copyright c 1996Elsevier Science Ltd
Printed in Great Britain. All rights reserved
0883-2927/95 $9.50+0.00

088%2927(95)00056-9

Soil sorption of caesium modelled by the Langmuir and Freundlich

isotherm equations
Linda S. Campbell
Department of Mineralogy, The Natural History Museum, Cromwell Road, London SW7 SBD, U.K.

and

B. E. Davies
Department of Environmental Science, University of Bradford, Bradford BD7 IDP, U.K.

(Received 13 April 1995; accepted in revised,form 16 July 1995)

Abstract-The Langmuir and Freundlich adsorption isotherm equations were used to model Cs sorption in
6 Welsh soils. The pH (in water), of the soils ranged from 3.5 to 6.5, and the estimated organic matter
(determined by loss-on-ignition) ranged from 4.8% to 46%. According to the Langmuir adsorption
isotherm, highest sorption maxima, b, were displayed by the loams at 102-204 mg kg-, compared with
organic soils at 37-86 mg kg-. Subsoil horizons (15-30 cm depth) did not always display higher sorption
maxima than their topsoil (O-l 5 cm depth) counterparts. Where an improved fit to the Freundlich model was
found over the Langmuir model, it was assumed that either concurrent desorption of other species, or bond-
energy heterogeneity in the substrate. were significant for these soils.

OBJECTIVES solutions in equilibrium with a known amount of

solid, hence, providing sorption determinations which
The behaviour of Cs in soils is of great significance in are theoretically analogous to those of gas-solid
relation to pollution from radioactive fallout events, systems.
such as the Chernobyl disaster of April, 1986. For The adjusted Langmuir equation, in its linear
many years after this event, the radioisotope 13Cs has form, is
persisted in a bioavailable form in many of the
organic, upland soils of Wales, Cumbria, and western C/(x/m) = 1/Kb + C/b (Bohn et al., 1985)
Scotland (Cawse et al., 1988; Beresford et al., 1992).
C (mg dm--?) is the concentration of adsorbate left in
However, knowledge of Cs geochemistry at the time
solution at equilibrium. K is the Langmuir bonding
suggested that it would either become fixed by
energy coefficient, b (mg kg-) is the adsorption-
exchange or adsorption mechanisms onto sheet
maximum, and x/?n (mg kg-) is the amount of
silicate minerals, or would be leached away, owing to
adsorbate adsorbed per unit mass of soil (or ad-
the high solubility of the alkali elements as singly
sorbent). The Langmuir isotherm is constructed by
charged cations. As part of a more detailed investiga-
plotting C/(x/m) against C; if the data lie on a straight
tion of Cs behaviour in soils (Campbell, 1994), this
line then the Langmuir model can be considered to be
study assesses whether the sorption behaviour of Cs
appropriate. Using least squares linear regression, the
can be modelled by the Langmuir or Freundlich
terms b (reciprocal of the slope) and K (( l/intercept)/
isotherms.
b), may be found.
A different model for adsorption, the Freundlich
isotherm, was also considered. Unlike the Langmuir
INTRODUCTION
model, differences in sorption enthalpy between
different types of site are accounted for by the
The Langmuir equation, originally formulated for
Freundlich model (Atkins, 1990). A recent mathema-
the determination of gas adsorption onto plane solid
tical derivation of the Freundlich equation, with notes
surfaces (Langmuir, 1918), requires the assumption
on its historical context and its applicability to surface
that the sorbate forms only a single layer. An account
heterogeneity, has been given by Sposito (1980). The
of the historical development of the equation is given
Freundlich equation is expressed as
by Harter and Baker (1977). The Langmuir isotherm
was adjusted by soil scientists to investigate solute
sorption. Essentially, replacement of the p (gas
pressure) term in Langmuirs original, to C (solution where 4 is the amount adsorbed per unit mass of
concentration), allows isotherms to be constructed for adsorbent, C is the solution concentration at equili-
715
716 L. S. Campbell and B. E. Davies

brium, and R and N are constants relating to filtration using 47 mm diameter, 0.45 btrn pore-size. cellulose
nitrate (CN) filter papers. This allowed a rapid separation of
adsorption capacity and intensity respectively (Pet-
the phases with potential adsorption of Cs onto apparatus
ruzzelli et al., 1985). If data conform to the Freundlich minimized.
model, a plot of log C versus log 9 should give a
straight line. (Note: 4 is equivalent to x/m in the
Langmuir plots.)
E.xperimental procedure

The essential procedure required to construct sorption

METHODS isotherms is to add known amounts of sorbate to sorbent,
equilibrate and measure the amount of sorbate left in
Materials and apparatus solution (Bohn et al., 1985). Replicate samples of the soils
were weighed into polyethene, screw-top tubes. Five con-
Six soil sampling locations were selected on the basis of centrations, approximately 2, 4, 6, 8, and 10 mg dmw3, of
altitude, parent litho!ogy, and soil type (Table 1). Two CsNO3 solutions were prepared for 10 sub-samples (includ-
locations were for typical upland organic soils (with their ing duplicates) for each soil-sample selected for the experi-
subsoil counterparts), and 4 were lowland loams, developed ment. The sealed tubes were shaken vigorously, and placed in
on a variety of substrates including Carboniferous Limestone an agitating water-bath in a cold room to equilibrate at 4C
(sample numbers 390~subsoil and 391-topsoil), Devonian for at least 16 h (the timing was pre-determined from a kinetic
Old Red Sandstone (1596, 1597), Lower Palaeozoic sedi- study; see below). This temperature was chosen because it
ments (401, topsoil only), and red till (1596, 1597). This set of could be kept relatively constant. Temperature fluctuations,
samples also provided a wide range in soil pH and estimated monitored regularly over a period of several hours, were
organic matter (loss on ignition, LOI%). Samples were taken always less than 0.X. At equilibration, the solutions were
by screw auger, bulked from 6 sub-samples taken from an filtered and aliquots were taken for dilution, acidification and
area of approximately 1 ml, and from a depth of O-15 cm addition of 1000 mg drne3 K (as nitrate) for analysis by
(topsoils) and 15-30 cm (subsoils). Before sieving through a 1 FAES (van Eysinga and van Elderen, 1988). The function of
mm mesh, the samples were air-dried at room temperature a large quantity of K in the solutions was to suppress
(approximately 20C). Total Cs was determined from each ionization of Cr., through the excess input of electrons from
sample through acid digestion after ashing, and analysis by K, which also easily ionizes to K.
flame atomic emission spectroscopy (FAES). To determine the time required for equilibration of the Cs
The CsNO, solutions were prepared from commercial solutions with the soils. a study was conducted for one
standard solutions by conventional serial dilution methods in upland soil and one lowland soil. The solution concentration
the same manner as for analytical standards. Separation of was5mgCsdm-3 and 1 g samples of the soils were agitated
the solutions from the soils after equilibration was by with 25 ml solution for specific lengths of time, ranging from

Table I. Characteristics of the soils used in the sorption experiments

Soil Location Altitude Type+ Drainage Parent PH Organic Natural Cs

with grid (metres, to and lithology (in water) matter content
reference nearest 10 m) (slope) (LOI%) (mg kg-)

363 Lake Vyrnwy 1013b Poor L. Palaeozoic 3.57 45.x 4.14

catchment 550 Crowdy 2 Mudstones
362 SH911231 Upland (Plateau) 3.53 8.2 8.17
peat
379 Hebron Stn. 611~ Reasonable L. Palaeozoic 3.92 35.6 4.02
SH 585585 310 Manod Muds and
378 Upland (Gentle) siltstones 4.22 4.8 5.36
podzol
391 Great Orme 313c Good Carboniferous 6.00 20.2 8.89
Head 120 Crwbin Limestone
390 SH 779836 Loamy (Steep) 6.49 8.9 8.42
ranker
1583 Trefnant I 1I,?, Poor Red till on 5.33 10.8 4.23
SJ 038704 8.5 Salop U. Palaeozoic
1582 Loamy (Flat/gentle) red sandstones 5.80 8.9 4.45
stagnagley
1597 R. Cilieni 541a Good Devonian 4.87 7.6 3.54
SN 927306 235 Milford Old Red
1596 Loamy red (Med./steep) Sandstone 4.96 6.4 2.37
brown earth
401 Farm near 541j Good Upper 5.00 II.7 3.99
Denbigh 140 Denbigh I Palaeozoic
SJ 033638 Brown earth (Gentle)

* Odd numbers = topsoils. Even numbers = subsoils.

+ Classification of the Soil Survey of England and Wales, with soil map code, Rudeforth er al. (1984).
 Soil sorption of caesium 717

30 min to 24 h. The filtration method of separating the for 4 data points. Similarly, the anomalous soil in the
solutions from the solids was rapid enough for precise timing
set of subsoils is number 378, corresponding to topsoil
to within a few minutes.
379.
Examination of the adsorption maxima (b, Table
3) suggests that most topsoils have lower capacities to
RESULTS
sorb Cs than their subsoil counterparts, but that soil
type is far more influential. Topsoil 379 again, is an
Sorption kinetics
exception, since it has an adsorption maximum nearly
twice that of subsoil 378. The soil types which display
The results of the preliminary kinetic study are
the highest adsorption maxima are the lowland loams
shown in Fig. 1. The most rapid increase in the
(390/391, 1582/1583, and 401) with pH values >5.
amount of Cs sorbed onto both soils occurred in the
Soils 1596/1597 have adsorption maxima closer to
first 4 h. At 8 h, the amount of sorption onto the
those of the upland soils than to the other loams.
upland soil was no different from the amount at 4 h
The Langmuir bonding energy coefficient (K) is
and, furthermore, did not change with any other
higher in the loams than in the upland soils, but again,
increase in time interval up to the maximum of 24 h.
K values in soils 1596/1597 are more characteristic of
The amount of sorption onto the lowland soil
soils 362/363 and 378 than of the other loams. Soil 379
however, continued to increase between 4 and 12 h,
is anomalous in its K value.
at a slower rate than at the start. Equilibration of the
Cs solution with the lowland soil was apparently
reached by 12 h. with no subsequent detectable change
Cs sorption: Freundlich isotherms
between 12 and 24 h.
Since the sorption kinetics for the 9 other soil
As with the Langmuir model, most of the soils
samples were unknown, a reaction time of at least 16
tested conformed to the Freundlich model of soil
h was selected on the basis of the longest measured
adsorption (Fig. 3 (a-f) and Table 4), with p&0.01,
equilibration time, plus 4 h extra time to allow for
indicating highly significant fits. Some of the soils were
some uncertainty. Sixteen h was also a convenient
better modelled by the Freundlich equation than by
time for practical reasons; allowing runs to be set up
the Langmuir equation but no consistent patterns of
overnight.
upland versus lowland, or topsoils versus subsoils
were detected.

Cs sorption: Langmuir isotherms

Cs sorption andpH
Selected Langmuir isotherm plots are displayed in
Fig. 2(a-f) from experimental results in Table 2. Changes in suspension pH after interaction of Cs
Regression data, and adsorption parameters obtained solutions with soils were monitored for various
from them, are shown in Table 3. All positive starting concentrations of Cs in solution. It was
correlation coefficients for the topsoil Langmuir necessary to prepare a set of runs in parallel to the
isotherms are highly significant at r20.99, and main sorption experiments in which Cs was to be
r > 0.98, except for soil 363 (r2 = 0.948) and soil 379
analyzed, since the pH measurements might have
which is significant at only the p < 0.05 level, and only contaminated the solutions. Actual starting concen-
trations were identical to those used in the main
sorption runs.
Figure 4 shows the effect of starting concentration
on final, equilibrium pH as affected by each soil. The
blank solutions (no interaction with soil), display the
lowest values of pH, ranging from 3.11 at 1.9 mg
dm- to 2.41 at 9.8 mg dmW3. All soils buffered the
E 60 solution pH to a certain extent, with the highest pH
values occurring in the lowland soils. In particular,
H
soils 390 and 391, associated with the Carboniferous
i 60
Peat sample 363 Limestone, appeared to buffer the solution pH most
8
-- 40 II ??

*-:;._~
...... .,~ . --* .__,
,,*. .........-.
--I-....-._.___..
I effectively. In all cases, the solution pH at equilibrium
decreased with increasing start concentration of cs

30 0 6 10 16 20 26
Time (hours) DISCUSSION

Fig. 1. Sorption kinetics of Cs for two topsoils; the Crowdy 2

upland peat (sample 363) and the Denbigh 1 loamy brown The Freundlich adsorption isotherm is generally
earth (sample 401), at 4C. regarded as being less sensitive than the Langmuir
718 L. S. Campbell and B. E. Davies

0
Y 4 ! cu 0
0 0 0 0 2
 Soil sorption of caesium 719

Table 2. Sorption data from which the Langmuir and Freundlich isotherms were modelled. All data are for Cs sorption from
aqueous solution

Topsoils Subsoils

Sample Total Equilibrium Sample Total Equilibrium

number sorbed concentration number sorbed concentration
(-+I (0 (x/m) (Cl
mg kg- pg drne3 mg kg- pegdm--3

363 24.5 18.3 362 33.2 14.8

40.3 54.9 48.1 48.7
50.0 95.3 60.8 88.4
55.5 135.6 68.4 128.2
44.9 174.9 73.9 167.6
379 25.5 17.2 378 7.9 35.1
41.1 53.5 16.5 79.1
57.7 88.7 20.1 125.2
53.3 139.8 20.5 171.9
39.7 181.5 4.1 217.5
391 41.2 1.9 390 41.7 1.1
84.9 11.7 87.0 7.9
115.2 26.0 122.1 21.9
145.0 48.0 151.2 44.2
147.0 70.3 154.4 62.6
1583 45.9 0.3 1582 46.3 0.3
88.1 5.7 92.9 4.4
126.2 19.4 131.6 15.1
158.7 37.6 169.3 33.7
176.8 62.4 195.2 45.4
1597 37.8 9.9 1596 35.7 10.9
52.8 44.5 51.4 47.4
63.6 82.6 61.7 83.9
78.5 119.4 71.4 127.2
72.7 171.9 74.0 170.6
401 37.7 4.6
72.8 21.9
94.3 51.0
102.8 89.2
92.9 125.5

.y/rnis the total amount ofCs adsorbed per unit mass soil. C is the concentration ofCs left in solution at equilibrium. Starting
concentrations of the Cs solutions were approximately 2,4,6,8, and 10 mg dm-.

Table 3. Langmuir isotherms: Linear regression data and Cs adsorption parameters for 11 soils at 4C

Correlation Significance Adsorption Bond-energy

coefficient (see key below) maximum coefticient
Soil r r-squared Nmg kg-) K

401 0.995 0.990 *** 102 0.158

363 0.974 0.948 ** 53.5 0.068
Topsoils 379 0.956 0.913 * 45.1 4.269
391 0.995 0.991 *** 167 0.114
1583 0.992 0.985 *** 185 0.193
1597 0.991 0.982 ** 81.3 0.060
362 0.995 0.990 *** 85.5 0.032
378 0.959 0.920 * 36.6 0.009
Subsoils 390 0.997 0.994 *** 169 0.167
1582 0.987 0.975 ** 204 0.216
1596 0.995 0.990 *** 82.6 0.045

Significance: pCO.05; pbO.01; p<O.OOl.

 ,,
720 L. S. Campbell and B. E. Davies

:.T : :
;..i
: 1. :

: j :

.... ..... ., . . . .. . . . ;.
: :

: :
j :

I :
i

..,.
 Soil sorption of caesium 721

Table 4. Freundlich isotherms. Linear regression data and Cs adsorption parameters for 11soils at 4C

Correlation Significance Freundlich constants

coefficient (see key below) (see text)
Soil r r-squared n K

401 0.989 0.978 ** 2.88 23.1

363 0.997 0.995 *** 2.41 7.44
Topsoils 379 0.748 0.560 3.88 13.82
391 0.995 0.990 *** 2.74 33.8
1583 0.997 0.994 *** 3.86 59.7
1597 0.975 0.950 ** 3.93 20.9
362 0.999 0.998 *** 2.99 13.42
378 0.961 0.923 * 1.59 0.91
Subsoils 390 0.995 0.991 *** 2.98 41.6
1582 0.998 0.997 *** 3.50 62.6
1596 0.997 0.994 *** 3.67 IS.45

Significance: *p<O.O5:??
*p<O.Ol; *p<O.OOl.

isotherm, in describing sorption phenomena (Bohn et Freundlich modelling, rendering the use of the
al., 1985). This is because more assumptions are equations in soil sorption studies the subject of much
required in the Langmuir model; specifically, that the contention over the last two decades. Harter and
substrate is homogeneous with respect to the bonding Baker (1977) pointed out that using the Langmuir
energy of sites (e.g., Petruzzelli et al., 1985), and that equation in solution studies in direct analogy to
no desorption takes place. Hence, where a soil fits the Langmuirs original for gaseous systems is potentially
Freundlich model better than the Langmuir model, as problematical. The reason is that when a species is
in soils 362/363, 1582/1583, and 1596, it is supposed sorbed from solution, there is accompanying desorp-
that these assumptions are not valid for these tion of another species, for charge-balance consider-
particular soils. Whether the sorbent is homogeneous ations. Another problem, as described by Veith and
or not with respect to the bonding energy, determina- Sposito (1977) is that removal of the solute species in
tion of the coefficient Kis questionable. Nychas (1984) question is not necessarily an indication of an
found that measured bond-energy coefficients were adsorption mechanism. Removal might be by pre-
dependent on the type of method used, and hence were cipitation, possibly, as suggested by Veith and Sposito
unreliable in their chemical significance. However, the (1977) explaining the sudden apparent increases in
sorption maxima, b, were found by Nychas (1984) not adsorption observed in the high concentration
to be dependent on experimental method. ranges of many experiments. For Cs, however,
Processes and mechanisms for sorption have been precipitation would seem unlikely owing to its high
widely considered in relation to Langmuir and aqueous solubility.

5.5-

64' SUBSOILS
TOPSOILS

2-1 I,,,,, I $ 1
2.6-1 I I I, I f I 012 3 4 5 6 7 6 9
012 3 4 6 6 7 6 0 10
Start solution(approx: mg Cs dm A -3)
Start solution (approx: mg Cs dm A -3)

Fig. 4. pH changes in reaction suspensions as a result of interactions of soils with CsNO, solutions. (a)
Topsoils, (b) subsoils.
122 L. S. Campbell and B. E. Davies

Some workers, for example, Singh ef al. (1987) and exchange capacities of these soils (not determined),
Erten et nl. (1988), concluded that 2-part fits for their and the absence of leaching forces (such as high
K sorption data were appropriate, suggesting different rainfall, high permeability, low pH, low rate of
sorption sites or processes. Although Charles and organic decomposition). However, soil 1596/7, resting
Prime (1983), purported that Cs is confined to edge on the Devonian Old Red Sandstone (ORS), contain-
sites of micaceous minerals, the possibility of ing a high amount of Fe oxides and quartz (sand
exchange of Cs onto interlayer sites has previously grains), did not display the capacity to sorb Cs any
been put forward by Elprince (1978) in examining the more than the upland soils. This suggests that Cs,
relationship of pH-dependent charge (relating to unlike other heavy metals, does not have an affinity to
hydroxy-Al equilibria), and Cs, in layer silicates. sorb onto, or co-precipitate with, Fe oxides.
Cation exchange phenomena might therefore explain Improvements noted in goodness of fit when the
the non-linearity of fit of other workers sorption data highest concentration level was omitted from some of
to the Langmuir or Freundlich models, indicating that the soil data, could be an indication that concentra-
precipitation of a Cs phase would be unnecessary. tion may not equate to activity of Cs at these levels. It
With regard to sorption kinetics, the equilibration could also indicate fundamental changes in the
of the Cs solutions of the present study within 12 h is substrate caused by ion exchange or by the lowering
consistent with the work of Sawhney (1966) who of pH. For example, Al solubility is known to be
noted that Cs sorption onto illite and montmorillonite highest at extremes of pH, and it is conceivable that
was rapid (a few hours). However, slow equilibration mineral-lattice breakdown began to occur, causing the
(> 3 weeks) of Cs solutions with vermiculite was opportunities for Cs sorption to decrease.
attributed by Sawhney (1966) to slow diffusion of Cs
into interlayer sites. This would again suggest either a
cation exchange phenomenon, or an absorptive CONCLUSIONS
process, rather than simple surface interaction. For
the purpose of determining adsorption kinetics there- With the exception of one upland locality, all soils
fore, the observation with vermiculite by Sawhney that were tested for Cs sorption conformed signifi-
(1966) would appear to be irrelevant. cantly either to the Langmuir model, or to the
The soil which least fitted the Langmuir model, soil Freundlich model, or to both. Where the Freundlich
3781379, is an upland podzol with a pH of 3.92 in the equation gave a better fit to the data, inappropriate
surface horizons. Additionally, this soil location had assumptions about the use of the Langmuir equation
the lowest values for parameter 6, the adsorption in a soil-solution context were indicated.
maximum, compared with all other soils in the The adsorption maximum (b in the Langmuir
experiment. Collectively, the data indicate that lea- equation), is accepted as a reasonable indication of
ched soils do not favour Cs sorption, and this may the sorption capacity of a soil with respect to the
partly be caused by the typically low pH values of element being studied. The high values of b found for
leached, upland soils. The organic peat, soils 362/363, loams and for subsoils relative to the more organic
also had low adsorption maxima, but not as low as the topsoils are consistent with the generally accepted
podzol. The work of Livens et al. (1991) helps to view that Cs sorption in soils is onto clays and micas.
explain the indirect effect of organic matter on the A lower value of b is consistent with the observed
ability of a soil to sorb Cs. In organic-rich soils, clay- higher bioavailability of Cs in upland soils, causing
organic complexation can occur, which effectively the continued radiocaesium contamination problem
blocks the clay sites that might have been available of upland vegetation and of upland sheep that graze
for Cs sorption. With majority opinion stating that the on it.
main type of Cs sorption is onto sheet silicate minerals
(e.g. Charles and Prime (1983) Lieser and Steinkopf Acknowledgements-This study was undertaken in the
(1989) and others in this discussion), the suggestion of Department of Environmental Science, University of Brad-
ford. L. S. Campbell was funded by the Welsh Office.
Livens et al. (1991) would appear to be central to the Improvements to the manuscript suggested by G. Cressey
understanding of Cs behaviour in upland soils. (The Natural History Museum), M. Graham and an
A wider study might highlight the contribution of anonymous reviewer are gratefully acknowledged.
the parent material to resultant adsorption para-
Editorial handling: Dr R. Fuge.
meters, since rock composition and grain size influ-
ence the pH and permeability of a soil, ultimately
affecting leachability (White, 1987). The preference of
REFERENCES
Cs for micas and illite over montmorillonite and
kaolinite is well documented (e.g. Singh and Gilkes
(1990); Komarneni (1978); Livens and Loveland Atkins, P. W. (1990) Physical Chemistry, 4th edn, pp. 995.
(1988)) and so the composition of a soil is also likely Oxford Univ. Press.
Beresford, N. A., Howard, B. J., Barnett, C. L. and Crout,
to influence its sorption parameters considerably.
N. M. J. (1992) The uptake by vegetation of Chernobyl
The high measured adsorption maxima in the high- and aged-radiocaesium in upland west Cumbria. J.
pH, lowland loams might be associated with the Environ. Radioactivity 16, 181-195.
 Soil sorption of caesium 723

Bohn, H., McNeal, B. and OConnor, G. (1985) Soil Livens, F. R. and Loveland, P. J. (1988) The influence of
Chemistry, pp. 341. Wiley. soil properties on the environmental mobility of Cs in
Campbell, L. S. (1994) Behaviour of caesium in soils: A Cumbria. Soil Use and Management 4, 69-75.
geochemical study with special reference to Wales. Nychas, A. E. (1984) On the determination of the constant
Unpublished Ph.D. thesis, University of Bradford. parameters of the Two-surface Langmuir equation.
Cawse, P. A., Baker, S. J. and Jenkins, D. (1988) A post- Zeitschrifr fur Planzenerntihrung und Bodenkunde 147,
Chernobyl survey of radionuclides in Wales, August- 584-591.
October 1986. Harwell Laboratory, Oxfordshire. AERE Petruzzelli, G., Guidi, G. and Lubrano, L. (1985) Ionic
R-12828, pp. 28. strength effect on heavy metal adsorption by soil.
Charles, D. and Prime, D. (1983) Desorption behaviour of Commun. Soil Sci. Plant Anal. 16, 971-986.
artificial radionuchdes sorbed onto estuarine silt: (I) Cs- Rudeforth, C. C., Hartnup, R., Lea, J. W., Thompson, T.
137 and Ru-106, (II) Zr-95 and Nb-95. J. Environ. Poll. R. E. and Wright, P. S. (1984) Soils and their use in
(Series B) 5, 273-295. Wales. Soil Survey of England and Wales, Bull. 11.
Elprince, A. M. (1978) Effect of pH on the adsorption of Sawhney, B. L. (1966) Kinetics of Cs sorption by clay
trace radioactive Cs by sediments. Water Resources Res. minerals. Soil Sci. Sot. Am. Proc. 30, 565-569.
14, 696698. Singh, B. and Gilkes, R. J. (1990) Sorption-desorption
Erten, H. N., Aksoyoglu, S., Hatipoglu, S. and Gokturk, H. behaviour of Cs in some West Australian soils. Australian
(1988) Sorption of Cs and Sr on montmorillonite and J. Soil Res. 28, 929-946.
kaolinite. Radiochim. Acta 44-45, 47-151. Singh, A. P., Mehta, S. C. and Mittal, S. B. (1987)
Harter, R. D. and Baker, D. E. (1977) Applications and Potassium adsorption in some potassium depleted soils.
misapplications of the Langmuir equation to soil adsorp- J. Indian Sot. Soil Sci. 35, 375-382.
tion phenomena. J. Soil Sci. Sot. Am. 41, 1077-1080. Sposito, G. (1980) Derivation of the Freundlich equation
Komarneni, S. (1978) Caesium sorption and desorption for ion exchange reactions in soils. J. Soil Sci. Sot. .4m.
behaviour of kaolinites. J. Soil Sci. Sot. Am. 42, 531-532. 44, 652-654.
Langmuir, I. (1918) The adsorption of gases on plane van Eysinga, I. P. N. L. and van Elderen, C. W. (1988)
surfaces of glass, mica and platinum. J. Am. Chem. Sot. Uptake of Cs by glasshouse vegetable crops from soil and
40, 1361-1403. nutrient solutions. Acra Horti. 222, 129-134.
Lieser, K. H. and Steinkopf, T. (1989) Chemistry of Veith, J. A. and Sposito, G. (1977) On the use of the
radioactive Cs in the hydrosphere and in the geosnhere. Langmuir equation in the interpretation of adsorption
Radiochim. Acla 46, 39-47. - phenomena. J. Soil Sci. Sot. Am. 41, 697-702.
Livens, F. R., Horrill, A. D. and Singleton, D. L. (1991) White, R. E. (1987) Introduction to the Principles and
Distribution of radionuclides in the soil-plant systems of Practice of Soil Science, 2nd edn, p. 244. Blackwell,
upland areas. Health Phys. 60, 539-545. Oxford.