Linear isotherms were obtained for the initial very low concentration range. Linear distribution coefficients (K
d
values) were obtained from the initial slope (first two data points) of the linear graph of c
s
against c
eq.
$$ {c_s} = {K_d}{c_{\text{eq}}} $$
(1)
-
c
s
… sorbed amount of substance on the solid surface (mg kg -1)
-
c
eq … solution equilibrium concentration (mg L -1)
-
K
d
… sorption (distribution) coefficient [(mg kg -1)/(mg L -1)]
-
K
oc = K
d
/(f
oc), with f
oc = fraction of organic carbon
At higher, i.e., more realistic concentrations, or when the concentration spans several orders of magnitude, the Freundlich equation is more suitable for fitting the sorption behavior: sorption isotherms were thus also analyzed using the Freundlich equation and represented in a double-logarithmic plot:
$$ {c_s} = {K_f}{c_{\text{eq}}}^{{1/n}} $$
(2)
-
in the linearized form: \( \log \,{c_s} = \log \,{K_f} + 1/n\,\log \,{c_{{eq}}} \)
-
K
f
… Freundlich constant [mg1 − 1/n kg -1 L1/n]; more intuitively: K
f
= c
s
at c
eq = 1 mg L−1
-
1/n … Freundlich exponent (slope of the double-log plot, or deviation from linearity in the linear form)
A typical shape of sorption isotherms is represented in Fig. 1 and discussed below.
The influence of pH on the adsorption of MCPA
Sorption experiments were performed at pH = 3 and at the unadjusted pH of ∼7. Upon acidification, an increase of sorption was observed; there is a clear shift of the sorption curves to the upper left (compare Fig.1 b with Fig.1 a, for both soils C and E, respectively). Freundlich parameters (K
f
and 1/n) and linear distribution coefficients K
d
as well as K
oc values are presented in Table 2. The K
oc is the sorption coefficient normalized to the organic carbon content; the values are not significantly different in most cases, except for the measurements at pH = 3, where the values are up to one order of magnitude larger.
Table 2 Freundlich sorption fits
The lipophilic aromatic part of MCPA is not influenced by the pH, in contrast to the carboxylic group. The pK
a
value of MCPA (pK
a
= 3.07) implies that the compound exists mainly in the deprotonated anionic form throughout a wide pH range in natural waters and soil solutions. The results obtained indicated the greatest sorbed amount of MCPA at pH = 3 (where ∼50 % are present as deprotonated (anionic) and protonated (neutral) species, respectively). Ionic interaction of oppositely charged species is stronger than Van der Waals interactions; this is especially important in the discussion of the present case, considering the pK
a
value of MCPA and the point of zero charge (pzc) of many of the soil components, e.g., iron oxides, montmorillonite, and illite: pzc > 8 (Scheffer and Schachtschabel 2002). At pH values between the pK
a
and pzc, the mainly present anionic solution species (deprotonated MCPA in the carboxylate form) strongly interact with positively charged surface sites, whereas charcoal offers mainly uncharged surface sites and preferably sorbs neutral solution species (e.g., the neutral, i.e., protonated MCPA). This trend of increasing sorption of acidic herbicides with decreasing pH, as expected from theory, was observed by many authors like Haberhauer et al. (2001), Paszko (2011), and Hiller et al. (2012); besides, Paszko (2011) distinguished between hydrophobic and hydrophilic sorption of MCPA. The surface charge of the present soils was not determined here but can be expected to be net positive, at pH values below the pzc of the predominant soil components (e.g., Iglesias et al. 2010).
Influence of soil composition and complexity
With increasing complexity of the soils (compare soils C vs. E and H after 3 months of pre-incubation, Fig. 1a and b, as well as Table 2), K
f
values increased and the linearity decreased, as shown by the 1/n values falling below 0.5. However, in this case, the K
d
values did not show a clear trend; the discrepancy between K
f
and K
d
values was also discussed by Karnjanapiboonwong et al. (2010). The charcoal content was expected to play a crucial role; there is a slight indication of its influence at both pH values in Fig. 1 (shift to the upper left, comparing soil C vs. E). The measured BET values are not that high (cf. new data in Table 1), so its influence is not expected to overrule any other contributions of other components. As a consequence, K
oc data did not show a clear trend, except for the drastically increased values at pH = 3.
The influence of biological aging of artificial soils on the sorption of MCPA
Typical sorption isotherms showed a shape like the one depicted in Fig. 2; additional information is given in Table 2. The obtained Freundlich and linear sorption fits are summarized for both pH values; for explanations cf. Eq. 1 and thereafter. The youngest artificial soil exhibits significant sorption at lowest concentrations; the most prominent difference seems to be a change in sorption characteristics upon aging.
Although the fit of the Freundlich isotherm is not always perfect—especially for soil H (3 months)—it still seems to be the most reasonable model, as all other sorption theories give worse fits (e.g., Langmuir); the linear sorption model is applicable only at low concentrations.
K
f
decreases with increased soil aging, with a concomitant increase in the Freundlich exponent 1/n, whereas, according to the linear sorption model, the sorption capacity at low concentrations increases (shown by the increasing K
d
value), with a substantial decrease at higher concentrations. Upon acidification, sorption capacity increases as well: both K
f
and K
d
values increase compared to their values at pH = 7, and at the same time the 1/n exponent does not show a clear trend.
The results of N2-BET measurements of each component of the soils are given in Table 1. It is obvious that the sum of the components would give a much larger specific surface area for sorption reactions than the incubated artificial soils (compare the larger values of SSA of MO, IL, and CH with the very small one of soils C, E, or H). The reduction of the surface area after incubation of the components might be due to interactions of reactive surface and therefore a blocking of reactive sites. Incubation of the artificial soils further reduced the reactivity of the surfaces for sorption with time, probably due to the interaction of the components and their associations, achieving stable micro-aggregates (Six et al. 2004). These micro-aggregates are the basis for the formation of macro-aggregates due to exudates of microorganisms like fungal hyphae or bacteria.
Measurement of sorption by FTIR (detection of surface species)
FTIR spectra of MCPA, soil H after 3- and 18-month incubation is shown in Fig. 3, along with the soils plus sorbed MCPA after a simulated sorption experiment: in the latter case, soil was impregnated with a solution of MCPA; the solution was not removed but evaporated. Subsequently, a KBr pellet with highly MCPA loaded soil sample was prepared. Nevertheless, the concentration was probably below the limit of detection to show any trace of residual or sorbed MCPA on the soil. In contrast to Pusino et al. (1995), the adsorbed herbicide could not be detected with FTIR. Yet, the most prominent features of the spectra of both pure MCPA and soils can be distinguished, i.e., the bands of carboxylic and carbonyl groups are located around 1,745 cm−1 (MCPA dimer) and 1,710 cm−1 (monomer), and 1,650 cm−1 in soils.