SSSAJ Grow Your Career with SSSA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 12 March 2007
Published in Soil Sci Soc Am J 71:280-288 (2007)
DOI: 10.2136/sssaj2006.0142
© 2007 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Hurrass, J.
Right arrow Articles by Schaumann, G. E.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Hurrass, J.
Right arrow Articles by Schaumann, G. E.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Hurrass, J.
Right arrow Articles by Schaumann, G. E.
Related Collections
Right arrow Soil Kinetics
Right arrow Soil Organic Matter
Right arrow Soil Hydrology
Right arrow Water repellency

SOIL PHYSICS

Hydration Kinetics of Wettable and Water-Repellent Soils

Julia Hurrass and Gabriele E. Schaumann*

Inst. of Science Organic and Environmental Chemistry, Universtiy Koblenz-Landau Universitaetsstr.1, D-56072 Koblenz, Germany

* Corresponding author (schaumann{at}uni-koblenz.de).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The hydration kinetics of soil organic matter (SOM) are influential factors for transport and sorption processes in soil. Nevertheless, our knowledge about wetting and swelling processes, which both control the overall hydration kinetics, is limited. In this study, we observed the hydration process of actually water-repellent and wettable soil samples by three independent methods. The rate of water uptake by water-repellent samples was distinctly lower than that of wettable samples when the water was supplied in the liquid phase, but was comparable when water was supplied in the gas phase. Gravimetric measurements of the water uptake and 1H nuclear magnetic resonance relaxometry showed that wetting of water-repellent soils may last up to 3 wk. This duration is distinctly longer than the water drop penetration time of the samples, which consequently only reflects the first wetting step of the soil surface and does not consider infiltration of larger amounts of water. Since the achievement of equilibrium conditions is a slow process, the hydration kinetics of SOM may control transport and sorption kinetics in water-repellent soils. Differences in the freezing, melting, and evaporation processes of soil water of actually water-repellent and wettable soil samples, observed by differential scanning calorimetry (DSC), are probably only a consequence of different water contents under field conditions. Consequently, the DSC measurements provided no evidence of a general difference in the way water binds in water-repellent and wettable soil samples.

Abbreviations: DSC, differential scanning calorimetry • NMR, nuclear magnetic resonance • SOM, soil organic matter • TG, thermogravimetry • WDPT, water drop penetration time


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Under field conditions, soil moisture contents frequently change. Many studies have shown that water content, hydration time, and the type of water binding strongly influence the sorption of organic substances in SOM (Unger et al., 1996; Altfelder et al., 1999; Berglöf et al., 2000; Borisover and Graber, 2004; Schaumann et al., 2004). Additionally, hydration kinetics undoubtedly are very important for plant growth. Nevertheless, many aspects of hydration and drying processes are still not understood. The first step of the hydration process of soil samples includes the wetting of mineral and organic soil components. After wetting the surfaces, the water is distributed within the pore volume. Possibly, swelling of SOM and clay minerals as well as hydration of salts may follow. If one of these processes is slower than the others, it will be the rate-limiting step of the overall hydration process. Both wetting and swelling of SOM may be slow depending on the properties of the amorphous organic soil matrix, while the penetration of water into the pore volume occurs immediately as soon as the pore walls are wettable. We define wetting as the first contact between water and soil surfaces including the resulting alterations of the surface properties. Since wetting requires liquid water, water vapor cannot wet unless it condenses.

Water vapor uptake of isolated humic and fulvic acids was previously analyzed (e.g., Shatemirov et al., 1972; Chen and Schnitzer, 1976; Chiou et al., 1988). In a relative humidity of 90%, Chen and Schnitzer (1976) determined water contents (related to dry mass) of 23% for humic acids and of 51% for fulvic acids at 25°C. For 100% relative humidity and 23°C, Chiou et al. (1988) calculated a comparable equilibrium water content of 25%. Few studies deal with hydration of SOM in unfractionated soil samples, however. It can be assumed that the wetting and swelling processes differ between extracted humic fractions and the naturally occurring organic soil matrix, where large macromolecules and the extractable fractions are intimately mixed. The network structure between the distinctly smaller humic and fulvic acid molecules (Scheffer and Schachtschabel, 1992; Piccolo, 2002) may be controlled by types of bonds and crosslinks that are different from those in the soil matrix, whose characteristics are responsible for the degree of swelling and retarding hydration. Where the uptake of water vapor to bulk soil samples was studied, mainly high-organic-content soils were used (e.g., Rutherford and Chiou, 1992; Robens and Wenzig, 1996). Based on sorption isotherms, Rutherford and Chiou (1992) determined saturation water contents (related to dry mass) of 37% for the SOM of a peat sample from the Everglades and 39% for the SOM of a muck sample from the Michigan State University Muck Research Farm. Contrary to the results of Rutherford and Chiou (1992), water sorption isotherms of a peat sample measured by Robens and Wenzig (1996) indicated pronounced swelling associated with water contents about 60% (related to dry mass) for a relative humidity of 100%. Studies dealing with water vapor sorption by mineral soils have usually focused on adsorption on the soil mineral matrix (Orchiston, 1953; Sharma et al., 1969; Stawinski, 1983; Chiou and Shoup, 1985; Rhue et al., 1988).

Under field conditions, the water uptake from the liquid phase, rather than from the gas phase, may represent the dominant process in soil. For solid SOM, limiting water contents between 300 and 500% (related to dry mass) are cited (Scheffer and Schachtschabel, 1992). Presumably, the water contents achieved for humous substances are distinctly higher if the supply is via the liquid phase rather than the gas phase. The differences between water adsorption via liquid and gas phases can be explained by different interfacial tensions between soil and water in the former and soil and humid air in the latter.

The main object of this study was to monitor the hydration process of water-repellent and wettable soil samples, i.e., samples with significantly different wetting rates of their SOM, to get an improved insight into the underlying mechanisms. Miyamoto et al. (1972) also aimed to detect water sorption characteristics that reflect soil wettability, but they did not find differences uniquely related to repellency in terms of the amount of adsorbed water, the shape of adsorption isotherms, the heat of adsorption, or the integral free energy of adsorption. Since they studied water vapor sorption, they concluded that the effects of wettability and repellency on the water uptake process may only occur at high water contents that are not reached by water adsorption from the gas phase. Based on this experience and to reproduce the moisture conditions in the field, we supplied water to actually water-repellent and actually wettable mineral soil samples via the liquid phase and compared that with water uptake from the vapor phase. Since it is known that hydration and drying processes in soil may last up to 3 wk (Waksmundzki and Staszczuk, 1983; Altfelder et al., 1999; Schaumann et al., 2004), we traced water uptake by the samples as a function of time during this period.

Our fundamental hypothesis was that, in water-repellent soil samples, not only the first wetting step, but the whole hydration process proceeds at a lower rate than in wettable samples if the water is supplied via the liquid phase. This hypothesis is based on the assumption of comparable interfacial properties of the outer SOM surfaces, which are wetted at the beginning of the hydration process, and the bulk organic soil matrix, which is subsequently penetrated by the water molecules. Gaseous water molecules are supposed to penetrate SOM without the need to overcome interfacial tension. Consequently, no differences between water vapor sorption of wettable and water-repellent soils may occur (Miyamoto et al., 1972). This hypothesis was investigated by measuring the gravimetric water contents of wettable, water-repellent, and organic-free reference soil samples during hydration from a water supply via the liquid and gas phases.

Since the way of water binding in soil changes during alterations of the wettability in the course of hydration (Schaumann et al., 2004, 2005), water in water-repellent soils is supposed to be bound differently to that in wettable soils. Thus, in addition to the bulk water, a part of the water may be enclosed in fine pores (Berezin et al., 1973; Pfeifer et al., 1985), bound as gel water in SOM (Quinn et al., 1988; Tsereteli and Smirnova, 1992), or as hydrate water (Nishinari et al., 1997; Ping et al., 2001). Differential scanning calorimetry measurements of the freezing, melting, and evaporation processes in soil samples in combination with 1H nuclear magnetic resonance (NMR) relaxometry analyses, performed by Schaumann et al. (2005) for the soil samples analyzed in this study, were applied to elucidate whether water is bound differently in water-repellent and wettable soil samples. Specifically, the portions of unfreezable water were used as an indicator for the mean bond strength of the soil water (Quinn et al., 1988; McBrierty et al., 1996; Nishinari et al., 1997). The 1H NMR relaxation times also reflect the mobility of water molecules within soil samples, depending on the sizes of the water-filled pores as well as on the degree of interactions between water and pore walls. Particularly, the determination of the amount of water enclosed in fine pores was used for the investigation of the hypothesis of different ways of water binding in wettable and water-repellent samples. To examine, whether or not differences in the water binding are only a consequence of the usually different moisture contents of water-repellent and wettable samples in the field, the DSC and NMR measurements were also performed at several points of time after adjusting the water contents of the samples.

Previous studies have shown that wetting as well as swelling kinetics of soil samples can be observed by 1H NMR relaxation measurements (Todoruk et al., 2003; Schaumann et al., 2005). The swelling process is linked to a shift of the peak positions in the relaxation time distributions, while the wetting process is reflected by slow changes of the water portions between different ranges of relaxation times due to water redistribution within the pore system (Todoruk et al., 2003; Schaumann et al., 2004, 2005). The 1H NMR relaxometry additionally differentiates between wettable and water-repellent samples. For wettable samples, the wetting of all pore types occurs directly after sample moistening, whereas for water-repellent samples, the time constants of the first-order wetting process are on the order of several days (Todoruk et al., 2003; Schaumann et al., 2005).

For the evaluation of the wetting kinetics, Schaumann et al. (2005) divided the 1H NMR relaxation time distributions into four pore type regions (Type I: 0–15 ms, Type II: 15–90 ms, Type III: 90–600 ms, Type IV: >600 ms). These pore types represent different pore sizes and different bond strengths of the water molecules in the pores. Based on the surface relaxivities given by Hinedi et al. (1997) for silica and assuming a cylindrical pore system, Schaumann et al. (2005) estimated that Pore Type I should mainly consist of intraparticle micropores <0.5 nm and interparticle mesopores (0.5–180 nm). The Pore Type IV represents the free water in macropores. Due to the influence of pore surface characteristics and pore geometry, it is not possible to assign unambiguous pore sizes to the measured relaxation times, but the development of the water distribution in the different pore types can be observed during the course of hydration (Schaumann et al., 2005).


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Soil Samples
All experiments were performed using soil samples from the research location Tiergarten (Berlin, Germany) of the research group INTERURBAN. This inner-city park, which was created between the 17th and 20th centuries, represents an example for the occurrence of water repellency in an anthropogenic location. Due to the anthropogenic impacts, as, e.g., the deposition of construction waste after the Second World War, the location shows a pronounced small-scale heterogeneity of soil properties. Wettability changes within a few centimeters from easily wettable regions to spots with strong water repellency. A comprehensive description of this site is given by Hurrass and Schaumann (2006) and Schaumann et al. (2005). To minimize differences in soil properties except for wettability, four samples, T1 to T4, were taken directly side by side (maximal 1-m distance between them) from the humous upper layer (10–30 cm). Samples T1 and T3 were strongly water-repellent, while Samples T2 and T4 were easily wettable (Table 1). Subsamples were used in both field-moist and air-dried states. Air drying was preformed by equilibrating the samples for at least 3 wk with air at a relative humidity of 31% at 20°C. Additionally, subsamples were ashed for 2 h at 550°C.


View this table:
[in this window]
[in a new window]

  		Table 1. Characterization of the soil samples T1 to T4. The soil texture did not vary among the four samples.

 
Moistening of the soil samples was effected by adding prescribed amounts of water to subsamples using a microsyringe. Small water droplets were distributed within the samples to effect a homogeneous wetting without an input of mechanical energy.

Figure 1 summarizes which subsamples were used for the different experiments. All experiments were performed in two replications.


Figure 1
View larger version (23K):
[in this window]
[in a new window]

  		Fig. 1. Use of the different subsamples for the experiments performed in this study (WC: water content, x 2: replication with two samples). All moistened samples were derived from the field-moist samples.

 
Gravimetric Water Uptake
Water uptake of Samples T1 and T2 was investigated by water sorption of air-dried subsamples and their ashed organic-free reference subsamples via the gas phase. For this purpose, water-saturated air (99% relative humidity) was routed through a closed Perspex sample chamber. Plastic containers (3 by 3 cm) containing 1.5 ± 0.2 g of the dispersed soil in thin layers (3-mm thickness) were placed in the sample chamber. The other method operated with a water supply via the liquid phase. A ceramic plate (0.2 MPa air-entry value, standard flow, Soilmoisture, Santa Barbara, CA) was placed in a water-saturated sand bed. Circular Perspex containers (4-cm diameter) were placed on the ceramic plate with paper filters at their bottoms, into which 2.5 ± 0.8 g of the subsamples was dispersed in thin layers (3-mm thickness). Due to the sample dispersing, the resulting packing densities were only around 0.7 ± 0.3 g cm–3 in these layers. Water was supplied to the samples at atmospheric pressure. For both methods, the gravimetric water contents of the samples were determined at intervals during the course of sample hydration at 20°C. To minimize the disturbance of the hydration process, the samples were removed only for 20 to 30 s from the ceramic plate or from the sample chamber for the determination of their water contents by weighing them. The moisture contents are expressed on a dry-mass basis, unless stated otherwise.

Phase Transitions during Water Uptake
The processes of freezing, melting, and evaporation of soil water were analyzed for the T1 and T2 samples in the field-moist state and after adjusting prescribed moisture contents in these samples (Fig. 1). On one hand, a subsample of T1 was moistened to the field-moist water content of Sample T2 (30.2%). On the other hand, subsamples of both T1 and T2 were moistened to a distinctly higher moisture content of 55.0%. During storage of these moistened samples at 20°C, subsamples were taken at intervals and used for the phase transition measurements.

Phase transitions were analyzed with a Mettler Toledo DSC 822e (Mettler Toledo, Gießen, Germany) with an intracooler and nitrogen as purge gas (80 mL min–1). Subsamples (20 to 30 mg) were placed into standard aluminum crucibles (40 µL). DSC measurements of freezing and melting processes were performed in closed crucibles. Samples were cooled from 25 to –30°C with a rate of 2 K min–1 and after that heated to 15°C with the same rate. Holes were punched in the crucible lids before the evaporation measurements. Water evaporation from samples was induced by heating from 25 to 180°C with a rate of 10 K min–1. Evaluation of phase transitions was performed using the proprietary software STARe (Mettler Toledo), and further characterization of peaks was made using Origin 6.1 (OriginLab, Northampton, MA, USA).

Thermogravimetry (TG) measurements were conducted with a Mettler Toledo STGA 851e thermogravimetric system (Mettler Toledo, Greifensee, Switzerland) under dry air (200 mL min–1). Subsamples (0.5–0.6 g) were placed into ceramic crucibles. A crucible filled with Al2O3 was used as a reference. Samples were heated at 5 K min–1 from 25 to 950°C. The proportional weight loss (derivative TG signal) was recorded as a function of temperature. Evaluation of TG curves was performed with Thermal Advantage Version 4.0 software (TA Instruments, Alzenau, Germany).

Nuclear Magnetic Resonance Study of the Wetting Kinetics
The NMR experiments have been discussed previously (Schaumann et al., 2005). Experimental details necessary for understanding the further evaluation and the comparison of the NMR results to those of the other methods of this study are summarized here.

Before measurement, field-moist subsamples of T3 and T4 were moistened in the NMR containers to a water content of 54%. They were stored at 20°C and, at several intervals, they were subjected to NMR measurements. The measurements were performed with a 2 MHz Relaxometer (Maran 2, Resonance Instruments, Witney, UK). The transversal relaxation time T2 was measured by the CPMG (Carr–Purcell–Meiboom–Gill) pulse sequence (Meiboom and Gill, 1958) with an interecho spacing of 150 µs. For each relaxation decay curve, 64 scans were recorded.

The proton relaxation time T in a pore system is composed of the surface relaxation time TS within the distance {lambda} of the pore walls and the bulk relaxation time TB (Hinedi et al., 1997):

Formula 1[1]
where V and S are the volume and the surface area of the water-filled pores, and S/V is related to the pore size. Due to different pore sizes in heterogeneous pore systems such as soil, a relaxation time distribution rather than one single relaxation time occurs.

According to the evaluation of Schaumann et al. (2005), the proportion of water belonging to each of the Pore Types I to IV (see above) was calculated by adding all amplitudes within the corresponding range of the relaxation time distribution.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Gravimetric Water Uptake
Figure 2 shows the water contents of the air-dried soil samples T1 and T2 and their organic-free references vs. hydration time. For water supply via both the liquid and gas phases, the untreated soil samples reached higher water contents than their reference samples. Via the liquid phase, the hydration process of the untreated samples took >2 wk and resulted in water contents >80%. Contrary to that, the hydration process via the gas phase was completed within a few days and resulted in water contents between 3 and 4%. This points to an incomplete penetration of water into all parts of the pore system due to the absence of wetting of the organic soil matrix by gaseous water molecules. Because of the low packing densities of the air-dried soil samples dispersed in the sample containers before beginning the measurements, both gravimetric methods resulted in water contents distinctly higher than those that may be reached under natural undisturbed conditions.


Figure 2
View larger version (35K):
[in this window]
[in a new window]

  		Fig. 2. Water uptake of the air-dried Samples T1 and T2 and their organic-free references (two replications for each sample). Shortly after starting the experiment, the water contents of all samples that were moistened via the liquid phase proceeded above the axis break, while the water contents of all samples that were moistened via the gas phase remained below the axis break.

 
If it is assumed that the mineral phase of an unashed sample adsorbs water to the same extent as an ashed reference sample, an estimate of the water contents reached in the SOM phase is possible:

Formula 2[2]
where {Theta} is the measured water content of the whole sample, XSOM and XMin are the proportions of SOM and the mineral phase (see Table 1; XMin = 1 – XSOM). Using the water content of the mineral reference sample {Theta}Min, the water content of the organic soil fraction {Theta}SOM can be calculated; however, this approach yields negative values for SOM moisture contents of both Samples T1 and T2 for all points of time. Probably, the sample ashing at 550°C caused changes at the mineral surfaces (Kaiser and Guggenberger, 2003) or it opened additional mineral sorption sites, resulting in an increased water uptake in comparison with mineral compounds in untreated samples. If it is assumed that all water is absorbed by SOM alone, then for the water supply via the liquid phase, the final moisture contents of SOM would be 1130 ± 30% for T1 and 1500 ± 300% for T2. For the water supply via the gas phase, the final moisture contents would be 43 ± 4% for T1 and 35 ± 1% for T2. These values are comparable with the equilibrium moisture contents determined by Rutherford and Chiou (1992) for water vapor sorption to SOM of peat and muck samples, and they are between the calculated values for fulvic and humic acids (Chen and Schnitzer, 1976; Chiou et al., 1988). Consequently, our results do not confirm the hypothesis of pronounced differences between the water uptake of isolated humic fractions and solid SOM in unfractionated soil samples for water vapor sorption; however, they support the assumption of distinctly higher water contents for water uptake from the liquid phase in comparison with water uptake via the gas phase.

To examine, if the hydration process can be described by first-order kinetics, exponential functions were fitted to the progressions of the water contents {Theta}:

Formula 3[3]

Table 2 lists the final water contents, A, the time constants, {tau}, and the R2 for the studied samples. With one exception (hydration of T1 via the gas phase, R2 = 0.49), the results for R2 support this assumption. The water uptake process for water supply via the liquid phase was significantly slower for the water-repellent Sample T1 than for the wettable Sample T2 (Table 2). Contrary to the other samples, after 3 wk, the hydration process was not complete for Sample T1 (Fig. 2).


View this table:
[in this window]
[in a new window]

  		Table 2. Final water contents, A, time constants, {tau}, and R2 of the exponential functions fitted to the water contents as functions of hydration time. The errors were calculated on the basis of the difference between the measurement replications.

 
The water uptake process of the untreated samples cannot be described as a sum of two exponential functions in the form of Eq. [3] for the mineral and the organic phase assuming that the time constant of the mineral phase is comparable to that of the mineral reference sample. This approach would account for parallel hydration of the organic and mineral soil components, but it does not imply processes in series. If a sum of two exponential functions is used to describe data points of water uptake of T1 and T2 from the liquid phase, then the fitting results indicate a dependence among the fitting parameters. This interdependence shows that the data points are sufficiently described with one exponential function. If a sum of two exponential functions nevertheless is fitted to the data, none of the calculated time constants agree with those of the reference samples. This also may be based on alterations of the mineral soil matrix due to sample ashing.

The comparison of the water uptake from the liquid phase of the water-repellent Sample T1 and the wettable Sample T2 indicates a very slow wetting as the rate-limiting factor in the hydration process of these samples. The equivalent hydration rates of the organic-free reference samples, which are significantly faster than those of the untreated samples (Table 2), suggest that this slow wetting process is linked to SOM. The fast hydration rates for water supply via the gas phase are comparable for the samples (Table 2) and indicate the absence of a slow wetting process of SOM by this phase. Presumably, gaseous water molecules can penetrate the SOM structure without applying the wetting enthalpy. Likewise, the fast wetting rates in combination with the low equilibrium water contents may be explained by an exclusive wetting of mineral surfaces, if water is supplied via vapor intrusion.

Phase Transitions during Water Uptake
For a better comparability of the phase transitions of the field-moist and moistened T1 and T2 subsamples, in Fig. 3 DSC heat flows are standardized to the sample water content. The main differences between the field-moist samples are the lower transition temperatures of all phase transitions in the water-repellent Sample T1 (0.5, 0.9, and 36°C onset temperature difference with T2 for freezing, melting, and evaporation, respectively) and the additional peak at 145°C after the main evaporation peak for T1. Thermogravimetry (Fig. 4 ) has shown that this additional peak is connected to a weight loss, corresponding to evaporation of a small amount of water at a temperature above the main evaporation temperature of this sample. It is unclear, however, whether water is really responsible for the TG weight loss at this temperature. The temperature difference of this peak between DSC and TG measurements (peak maximum at 145 ± 1°C in the DSC thermogram and at 133 ± 1°C in the TG thermogram) may be caused by the different heating rates.


Figure 3
View larger version (46K):
[in this window]
[in a new window]

  		Fig. 3. Freezing and melting process (left) and water evaporation (right) of T1 after moistening to water contents of 30.2% (top) and 55.0% (middle) and of T2 after moistening to a water content of 55.0% (bottom). In all graphs, the field-moist T1 and T2 samples are included for comparison. The arrow pointing to the evaporation curve of the field-moist T1 sample indicates an additional evaporation peak.

 

Figure 4
View larger version (26K):
[in this window]
[in a new window]

  		Fig. 4. Thermogravimetry of the air-dried Samples T1 and T2. The arrow pointing to the thermogram of the T1 sample indicates an additional peak at 133 ± 1°C.

 
The specific freezing and melting enthalpies, i.e., the total energies under the peaks related to the water mass, are distinctly smaller for the field-moist Sample T1 than for T2, indicating that a fraction of water in this sample is unfreezable. Probably due to tight bonding (Quinn et al., 1988; McBrierty et al., 1996; Nishinari et al., 1997), this portion of water remained unfrozen despite cooling to –30°C at 2 K min–1.

With increasing water content, the freezing exotherms became broader and showed a composite structure in the form of resolved component peaks for both samples (Fig. 3, left), presumably caused by different degrees of mobility of the water molecules. After moistening, the melting temperature, the specific enthalpy, and the profile of the melting peak of Sample T1 approached that of T2. The melting onset temperature of T1 increased from –0.4 ± 0.1 to 0.0 ± 0.1°C for moistening to 30.2%, and to 0.1 ± 0.1°C for moistening to 55.0% water content. The error of 0.1°C in the melting onset is calculated on the basis of repeated measurements of subsamples, which were performed for different soil samples. For the freezing onset, the error is higher (1.5°C) due to the heat evolution of the exothermic freezing process, which counteracts the cooling by the DSC temperature protocol, resulting in nonreproducible temperatures in the samples. Consequently, contrary to the melting temperature, we could not observe that the freezing temperature of T1 significantly approached that of T2.

The slope of the linear regression in Fig. 5 shows, for a larger sample collection, that freezing enthalpies of soil water were generally smaller than the heat of fusion of free water, which is 333.5 kJ kg–1 (Angenheister, 1982). Because of higher proportions of unfreezable water, the water-repellent samples exhibited lower specific enthalpies than the wettable ones (Fig. 5), which supports the hypothesized stronger bonding of water molecules in water-repellent samples than in wettable samples. The absolute values of the specific freezing enthalpies of moistened samples were higher than those of field-moist samples (Fig. 5), indicating a higher amount of free or loosely bound water just after sample moistening.


Figure 5
View larger version (27K):
[in this window]
[in a new window]

  		Fig. 5. Freezing and melting enthalpies (related to wet sample mass) of field-moist samples from the locations Tiergarten and Buch (Hurrass and Schaumann, 2006; Täumer et al., 2005) as a function of the water content (related to wet sample mass). Additionally, the moistened T1 and T2 samples are included in the graph (15 min, 2 d, 7 d, and 21 d after moistening). All field-moist water-repellent samples are circled. The linear regressions were performed only for the field-moist samples.

 
The water content of 1.3 ± 0.6% (related to wet sample mass), which is linked to freezing and the melting enthalpies of 0 kJ kg–1 wet sample, corresponds with the intersection of the two regression lines (Fig. 5). This water content may serve as a measure for the amount of unfreezable water in the field-moist samples from the two considered locations.

After samples had been moistened, the specific freezing enthalpies decreased with time (Fig. 6 ), which may be caused by a slow binding of added water to the soil matrix. Due to the higher water contents after moistening, the low enthalpies of the original field-moist samples were not reached again. Comparison with a sand–water mixture (Fig. 6) indicates that SOM in the studied sandy soils is most probably responsible for the low specific freezing enthalpies associated with bound water.


Figure 6
View larger version (17K):
[in this window]
[in a new window]

  		Fig. 6. Specific freezing enthalpies of the moistened Samples T1 and T2 during the course of hydration, of the field-moist samples T1 and T2, and of a quartz sand–water mixture (19.0% water content related to dry sample mass). The errors were calculated on the basis of the difference between the measurement replications.

 
The variations of the specific freezing enthalpies can also be interpreted in another way: In the field-moist state, the smaller freezing enthalpy of water in T1 than that in T2 may be based on a higher salt concentration in the soil water of T1, which possibly is a consequence of a minor extent of leaching. Accordingly, the dilution of soil solutions following the abrupt increase in the water contents in both samples may have caused the rise in freezing enthalpies (Fig. 6). Then, in the course of time, a slow dissolution of additional ions could have produced a slow decrease in specific enthalpies (Fig. 6). In this context, the similar development of the specific freezing enthalpies of both Samples T1 and T2 after moistening (Fig. 6) may be explained by comparable variations in pH and electrolyte content, which are interrelated with wetting behavior in the studied location (Bayer and Schaumann, 2007; Hurrass and Schaumann, 2006). If ions are really released during hydration, presumably water is further penetrating the solid soil matrix, which would imply an increase in bond strength between the water and solid phases, so that this factor of influence additionally may affect the freezing enthalpies.

After 2 d of hydration, the freezing enthalpies of both samples moistened to 55.0% water content were comparable (Fig. 6). The same was observed for the onset temperatures of the phase transitions. Like the melting onset, the onset of the main evaporation peak of T1 also approached that of T2 and shifted from 64 ± 1 to 84 ± 1 and to 96 ± 1°C after moistening to 30.2 and to 55.0% water content, respectively (Fig. 3, right). Additionally, the area under the additional peak at 145°C of T1 decreased after sample moistening. Consequently, the equalization of the water contents of Samples T1 and T2 caused their phase transition characteristics to approach each other. Assuming that the phase transitions mainly reflect the bond strength of water and that the salt concentration represents only a minor factor of influence, the adjustment of the water contents results in comparable mobility of the water molecules due to a similar degree of interactions with the solid soil phase.

The phase transitions of soil water point to the existence of differently bound water molecules in the soil samples. A part of the soil water probably is bound very tightly to the SOM, so that it is unfreezable. In the field-moist state, this water proportion was higher for T1 than for T2, which supports our hypothesis of different mean bond strengths of water in water-repellent and wettable soil samples. The hypothesis also is supported by the additional peak at 145°C, presumably based on delayed evaporation of part of the water, and the lower freezing and melting temperatures of T1 (Berezin et al., 1973; Pfeifer et al., 1985; Nishinari et al., 1997; McBrierty et al., 1999; Ping et al., 2001). By adjusting the water contents of the samples, however, the differences in the phase transition characteristics could be eliminated within 21 d. Consequently, different bond strengths of water molecules in the field-moist samples are probably based on different proportions of water to soil and not on the different wettabilities. After moistening the samples, the added water perhaps had a higher mobility than that originally present, as reflected by the composite structure of freezing exotherms. The subsequent decrease of freezing enthalpy in the course of contact indicates a slow binding process of added water to the soil matrix. Since this process did not differ between T1 and T2, it probably does not reflect a wetting process, but possibly indicates a slow swelling process of SOM, which may occur in both samples irrespective of their wetting properties (Schaumann et al., 2005).

Nuclear Magnetic Resonance Study of the Wetting Kinetics
Contrary to the wettable Sample T4, the NMR relaxation time distributions of the water-repellent Sample T3 showed a high portion of free water (Type IV) directly after moistening (Schaumann et al., 2005). In Sample T3, the remaining water apart from the bulk water (Types I–III) revealed relaxation times that are, on average, smaller than those of T4 (mean relaxation time below 200 ms of 7 ± 2 ms for T3 and 32 ± 2 ms for T4; Schaumann et al., 2005). This may reflect stronger mean bond strengths of the water molecules that were already existent in the field-moist samples before moistening in the water-repellent sample than in the wettable sample, as also indicated by the phase transition characteristics. Thus, the NMR results support the assumption that the differences in the specific freezing enthalpies between the field-moist samples may be based on different degrees of water binding and not on different ion concentrations. Otherwise, the concentration of paramagnetic substances reducing the relaxation times also had to be significantly higher in T3 than in T4. This would be the case if differences in electrolyte content are directly linked with differences in Fe3+ and Mn2+ concentrations. Differences in the pore size distribution or the composition of the mineral matrix, however, likewise can explain the different mean relaxation times in Pore Types I to III.

Contrary to the conjectures deduced from the comparable development of the freezing enthalpies of the wettable and the water-repellent samples during hydration, no swelling process of SOM was observed by the NMR measurements (Schaumann et al., 2005). The study of Schaumann et al. (2005), however, has shown that during the wetting process of the water-repellent Sample T3, the free bulk water (Type IV) slowly moved to the smaller pores of Type I and II. The medium-sized pores of Type III acted as intermediate storage in this process with first-order kinetics, whereby the decrease of the water portion in Pore Type III yielded the highest time constant of 16 ± 4 d (Schaumann et al., 2005).

For a more comprehensive understanding of the wetting process, the NMR results (Schaumann et al., 2005) were compared with the results of the freezing enthalpies. Because of the direct proximity of the samples in the field and their similar properties (Table 1), a comparability of the water-repellent Samples T1 and T3 and of the wettable Samples T2 and T4 can be assumed.

By means of the freezing enthalpies of soil water (Fig. 6) and the heat of fusion of free water of 333.5 kJ kg–1 (Angenheister, 1982), the fraction of unfreezable water was determined and related to dry sample mass. In both Samples T1 and T2, the proportion of unfreezable water abruptly decreased after sample moistening (from 65 ± 3 to 5 ± 3% in T1 and from 36 ± 3 to 2 ± 3% in T2). In the course of hydration, the amount of unfreezable water increased from 0.02 ± 0.01 to 0.11 ± 0.01 kg kg–1 dry mass in T1 and from 0.01 ± 0.01 to 0.11 ± 0.01 kg kg–1 dry mass in T2.

For both the water-repellent and the wettable samples, the development of the amount of unfreezable water and of water in Pore Type I during hydration shows differences (Fig. 7 ). In the wettable sample, the water immediately wetted the pores of Type I, so that the amount of water in these pores did not change any more in the subsequent time steps. The portion of unfreezable water, however, only reached a constant value after 2 d of hydration time (Fig. 7, bottom). While the portion of water in Pore Type I gradually increased after moistening the water-repellent sample, the amount of unfreezable water rose abruptly between 2 and 7 d of hydration time (Fig. 7, top). Compared with the slow increase of water in Pore Type I, which is directly linked to alterations of surface properties of the inner pore walls, the amount of unfreezable water in the water-repellent sample reflects the wetting process only to a small degree. Due to hydrophobic properties of the pore walls, at first the water penetrated the coarse pores only. As a result of contact with water, the surface properties changed gradually, allowing the water to slowly penetrate finer pores as well. At the same time, the proportion of unfreezable water increased. But, the differences between the development of the freezing and the NMR data indicate that the unfreezability of a part of the soil water cannot be explained solely by its inclusion into fine pores.


Figure 7
View larger version (50K):
[in this window]
[in a new window]

  		Fig. 7. Comparison of the amounts of unfreezable water and of water in Pore Type I (Schaumann et al., 2005) during the course of hydration for the water-repellent samples T1 and T3 and the wettable Samples T2 and T4 after moistening them to 54 or 55% for the nuclear magnetic resonance and the differential scanning calorimetry measurements, respectively. The errors were calculated on the basis of the difference between the measurement replications.

 
In the wettable and the water-repellent samples, the amounts of water in Pore Type I were higher for all points of time than those of unfreezable water (Fig. 7). The presented differences even may be too small, because of a possible overestimation of the percentage of unfreezable water by the applied method (Kuntz and Kauzmann, 1974; Quinn et al., 1988). A part of the difference between unfreezable water and water in Pore Type I may be attributed to the use of different samples. The soil properties of both water-repellent and both wettable samples are so similar (Table 1), however, that their influence can probably be neglected.

Processes Involved in Soil Hydration
According to the NMR results, initially, the walls of Pore Types I to III of the water-repellent samples reveal hydrophobic characteristics, generally ascribed to the organic soil matrix (Wallis and Horne, 1992; Doerr et al., 2000). After moistening the samples, water molecules can gradually penetrate finer pores and, with that, alter their interfacial tensions. This slow change of the wetting behavior of the pore walls may be due to water-induced conformational changes of SOM, as the reorientation of hydrophobic and hydrophilic molecule groups (Ma'shum and Farmer, 1985; Valat et al., 1991; Roy et al., 2000). The development of unfreezable water proceeds simultaneously to the penetration of water into small pore types. Presumably, the inclusion of water molecules into fine pores is partially responsible for the increase in unfreezable water in the course of hydration (Berezin et al., 1973; Pfeifer et al., 1985). Other mechanisms, however, such as chemical binding to the solid soil matrix, can also contribute to a decreased mobility and consequently to an unfreezable status of a part of the water molecules (Quinn et al., 1988; Tsereteli and Smirnova, 1992; Nishinari et al., 1997; Ping et al., 2001). The differences partially occurring between the development of unfreezable water and the amount of water in small pore types during hydration (Fig. 7) support the existence of different mechanisms for water unfreezability in soil.

For the studied water-repellent samples, the determined time constants of the gravimetric water uptake and the redistribution of water molecules from medium-sized (Type III) to small-sized pores (Types I and II), measured by NMR, are comparable. They indicate that the wetting process may last 2 to 3 wk. Consequently, the distinctly lower water drop penetration times (WDPT; Table 1) reflect only the first wetting step of the outer soil surfaces.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
The hydration process of soil samples was characterized by three independent methods. The results indicate that water uptake is mainly influenced by SOM and that equilibrium water contents of the organic matrix are comparable for whole-soil samples and isolated humic fractions. A differentiation between water-repellent and wettable soil samples was only possible by direct measurements of the gravimetric water uptake if water was supplied via the liquid phase. Both 1H NMR relaxation analyses (Schaumann et al., 2005) and gravimetric measurements yielded first-order wetting kinetics of water-repellent soils, which lasted up to 3 wk. According to our results of the phase transition characteristics, differences in the bond strength of water molecules in wettable and water-repellent soil samples are probably a result of different moisture contents under field conditions.

For modeling transport and sorption processes, the hydration kinetics, which may be very slow in water-repellent soils, have to be taken into account. In some water-repellent soil regions, equilibrium conditions are hardly achieved, because the wetting process may last several weeks, rather than only some hours, as indicated by WDPT measurements. Consequently, in such soil regions, only time-delayed infiltration and subsequent water storage are possible after precipitation events. By contrast, equilibrium between water and the solid soil phase may be reached immediately in wettable soil regions if no swelling processes are involved. Substances that are transported with the water phase thus may interact instantly with wettable regions, while the penetration into water-repellent parts of the soil is retarded or may not occur at all.


   ACKNOWLEDGMENTS
 
We thank Jeannette Regnery for the realization of many DSC measurements and the DFG (project SCHA 849/4) for financial support.

Received for publication April 7, 2006.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Hurrass, J.
Right arrow Articles by Schaumann, G. E.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Hurrass, J.
Right arrow Articles by Schaumann, G. E.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Hurrass, J.
Right arrow Articles by Schaumann, G. E.
Related Collections
Right arrow Soil Kinetics
Right arrow Soil Organic Matter
Right arrow Soil Hydrology
Right arrow Water repellency

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome