Verification of the Solid–Liquid Separation of Waterlogged Reduced Soil via a Centrifugal Filtration Method

Abstract

The efficient separation of solid and liquid phases of soil under reductive conditions is of the utmost importance to study soil chemistry and to predict the mobility and bioavailability of nutrients and toxic contaminants in waterlogged reduced soils (WRSs). However, there is no established method for efficiently separating the solid and liquid phases of WRS within a short time while maintaining its reductive conditions. This study aimed to verify the applicability of a simple centrifugal filtration method (CFM) for the efficient separation of solid and liquid phases of a WRS and examine the CFM-extracted soil solution to confirm that the reductive condition was maintained during the solid–liquid separation process. Incubation experiments were performed under reductive conditions with or without ethanol/molasses used as additional organic material (OM), while the soil solution was collected by both a suction method and CFM at different centrifugation speeds (700, 2760, and 11,000 rpm) and times (1–7 min). The results showed that the soil pH increased with time while the Eh decreased, indicating that its reducing state was enhanced during the incubation experiments. The addition of OM promoted the reductive conditions in the first days of the experiments. Centrifugation speed, rather than time, was found to be the key to extract the maximum amount of soil solution, while a higher centrifugation speed (11,000 rpm), which represents the permanent wilting point, was found to be most effective for extracting the maximum amount of soil solution. The results exhibited no significant difference in solute (As, Fe(II), and Mn) concentrations when varying amounts of CFM-extracted soil solution were measured. The statistical analysis also indicated no significant (p > 0.05) difference between the solute concentrations in the CFM-extracted soil solution and the solute concentrations in the soil solution extracted by the suction method, confirming that the reductive condition was maintained during solid–liquid separation by CFM. This study suggests that CFM operating at a higher centrifugation speed could potentially be employed as a simple and highly effective technique to efficiently separate the solid and liquid phases of WRS (sandy clay loam) within a short time while maintaining its reductive conditions.

1. Introduction

The extraction and study of soil solutions is the basis for understanding pedogenic processes, the fate and transport of solutes, nutrient cycling, equilibrium and kinetic factors, and the bioavailability and leaching of toxic contaminants under both natural and engineered conditions [1,2]. Thus, the efficient separation of solid (soil) and liquid (soil solution) phases from the soil matrix is of paramount importance. Various in situ (field-based) and ex situ (laboratory-based) methods have been established to collect soil solutions, either with or without the separation of the solid–liquid phases of the soil matrix, under aerobic/upland conditions. For example, Elkhatib et al. reported that centrifugation is a rapid, reliable, and cost-effective method for soil solution collection and subsequent chemical characterization [3]. Dahlgren found that high-speed centrifugation was the simplest method that resulted in higher solute concentrations compared to column equilibrium or immiscible displacement techniques when the soil was sufficiently moist and investigated them to study pH-sensitive solutes such as trace metals [1]. Similar results were also reported by Geibe et al. when the authors compared the centrifugation drainage method with suction and zero-tension lysimeters [2]. Fraters et al. also observed constant concentrations of solutes with an increasing amount of soil solution when using the centrifugation drainage method [4].

On the other hand, the suction method has also been widely used to extract soil solutions, even from flooded lowland soils or waterlogged reduced soils (WRSs), which have been found to be associated with several problems resulting from soil macropore inhomogeneities and the spatial variability of the investigated soil properties [5]. Although studies reported that the pressure (or squeezing) method can be used to separate the water from sediment samples under a N₂ supply, this method requires several hours for such solid–liquid separation [6,7]. However, to the best of our knowledge, there is no established method for collecting soil solutions by efficiently separating the solid and liquid phases of WRS within a short time while maintaining its reductive (anaerobic) conditions.

Under waterlogged reduced conditions, significant changes occur in soil physicochemical properties, behavior and transport of essential nutrients, and toxic contaminants [8]. Among the major changes that occur are the alteration of soil pH, the lowering of the soil redox potential (Eh), the reduction of Fe(III) to Fe(II) and Mn(IV) to Mn(II), the reduction of nitrate (NO₃⁻) and sulphate (SO₄²⁻), an increase in the concentration and availability of P, Ca, Mg, Mo, and Si, and a decrease in the concentration and availability of Zn and Cu [8]. The reduction of As(V) to As(III) also occurs as a result of the reduction of Fe(III), Mn(IV), and SO₄²⁻ [9]. Therefore, the study of subsurface chemistry under waterlogged reducing conditions is of the utmost importance for a better understanding of the behavior of soil nutrients and trace metals in WRS, which is still scarce in the literature. Moreover, since rice (Oryza sativa L.), a staple food for about 3 billion people worldwide [10], is predominantly grown under lowland/waterlogged conditions where toxic metals (e.g., As, Cd, Cu, Pb and Zn) are the main contaminants of concern [8,11]. Acquiring in-depth knowledge on the dynamics of these toxic metals in soil solutions and solid phases under waterlogged conditions is of great importance for developing effective strategies to address and tackle the problem of toxic metal pollution in paddy soils.

Therefore, considering the advantages of the centrifugation method, including its ease of use, effectiveness, applicability to almost all soil types, and the benefit of reusing the separated soil for further studies [4], this study aimed to verify the applicability of a centrifugal filtration method (CFM) for the efficient separation of solid and liquid phases of WRS while maintaining its reductive conditions. Specifically, this research investigated soil pH, Eh, and solute (As, Fe(II), and Mn) concentrations in the soil solution extracted by CFM at different centrifugation speeds and times, while maintaining its reductive conditions, and compared these results with those obtained by the suction method.

2. Materials and Methods

2.1. Soil

Soil samples were collected from the plough layer (Ap horizon, at a depth of 0–15 cm) of a paddy field. Samples were collected from five different points in the field and then mixed properly to form a composite sample. The collected soil sample was then air dried at 25 °C, passed through a 2 mm stainless steel sieve, and stored at −4 °C until used for analysis and incubation experiments. Since the main purpose of our study was to compare the results obtained by suction method and CFM, air-dried and disaggregated soil was used in this study to start the incubation experiments, as this soil sample has uniform soil physicochemical properties.

2.2. Analysis of Soil Physicochemical Properties

The particle size distribution (% sand, silt, and clay) of the soil was determined using the pipette method [12]. Soil pH was analyzed using a digital pH meter (LAQUAtwin-pH-33, HORIBA Scientific, Kyoto, Japan) at a ratio of soil to either water or 1 mol/L KCl of 1:2.5 [13], while electrical conductivity (EC) was measured using a digital EC meter (D-52, HORIBA Scientific, Kyoto, Japan) in a soil–water suspension of 1:5 [14]. Total carbon (TC) and nitrogen (TN) contents were analyzed by the dry combustion method (Sumigraph NC-900, Shimadzu, Kyoto, Japan) [13]. The C/N ratio was calculated by dividing TC by TN. Exchangeable cations (Mg²⁺, Ca²⁺, K⁺, and Na⁺) extracted by ammonium acetate were determined by an inductively coupled plasma optical emission spectrometer (ICP-OES; 5110, Agilent Technologies, Santa Clara, CA, USA) [13]. The cation exchange capacity (CEC) of the soil was determined by displacement with NH₄OAc, while the base saturation percentage (BSP) was measured and expressed as the percentage of total soil CEC contributed to by Mg²⁺, Ca²⁺, K⁺, and Na⁺ [15]. The available phosphate (P₂O₅) was extracted and determined using Olsen’s extractant (0.5 M NaHCO₃ at pH 8.5) and method [16]. Acid ammonium oxalate-extractable Fe (Fe_o), Al (Al_o), and Si (Si_o) were extracted according to the methods described by McKeague and Day [17]. Sodium pyrophosphate-extractable Fe (Fe_p) and Al (Al_p) were extracted by the methods as described by Schuppli et al. [18], while dithionite-citrate-bicarbonate (DCB)-extractable Fe (Fe_d) and Al (Al_d) were extracted according to the methods described by Mehra and Jackson [19]. The concentration of metals in the extracts was determined by ICP-OES (5110, Agilent Technologies, Santa Clara, CA, USA) [20]. The concentration of 1 M HCl-extractable As was measured by inductively coupled plasma mass spectrometry (ICP-MS; ELAN^® DRC-e, PerkinElmer, Waltham, MA, USA). The DRC-e mode can effectively suppress the formation of a 40Ar + 35Cl complex. The blank samples were also measured with the same dilution ratio (100 times dilution) and the intensity of the blank samples was very low.

2.3. Incubation Experiments

Incubation experiments were carried out under flooded conditions to create a reducing environment, with or without additional organic materials (OMs). In this study, ethanol and molasses (a sugar-rich by-product of sugar extraction from sugarcane) were used as additional OM to enhance the reductive conditions in the soil [21]. For each treatment, 600 g of soil sample (oven-dried basis) was thoroughly mixed with ultra-pure water (w/v = 1/2, using a hand mixer fitted with an agitator blade, at 1180 rpm for 2 min), 0.5% ethanol, or 0.3% molasses. Each mixture was then placed in a 2 L glass beaker, and two silver chloride reference electrodes (LAQUA Electrode, HORIBA Scientific, Kyoto, Japan) and two soil solution samplers (DIK-300A, Daiki Rika Kogyo Co., Ltd., Saitama, Japan) were installed below the soil surface in each beaker. Thus, a total of 3 glass beakers were used to represent 3 different treatments, and the beakers were finally incubated for 18 days at 30 °C in an incubator (LH-220SP, Nippon Medical & Chemical Instruments Co., Ltd., Osaka, Japan). The reference electrodes were used to regularly monitor the Eh using a digital multimeter (LAQUAact-pH/ORP Meter-D-72, HORIBA Scientific, Kyoto, Japan), whereas the soil solution samplers (SSSs) were used to extract soil solutions (known as the suction method). The extracted soil solution samples were preserved using 10% HNO₃ or 1.19 M acetate solution for chemical analysis. The incubation experiments were conducted in duplicate.

2.4. Solid–Liquid Separation by CFM

For solid–liquid separation under reductive conditions, the surface water, electrodes, and SSSs were first removed from the incubated beakers at the end of the incubation experiments. The beakers were then quickly taken into a glove box (AS-800S, AS ONE Corporation, Osaka, Japan) and a thin layer (~1 cm) of the surface of the wet soil was removed under a continuous supply of N₂ while maintaining 0.1–1% O₂, indicated by an oximeter to prevent oxidation of the sample. The remaining wet soil was collected and centrifuged (CR20G, Hitachi Co., Ltd., Tokyo, Japan) at different centrifugation speeds (700, 2760, and 11,000 rpm) and times (1–7 min), with 90 g of soil taken for each centrifugation. Three different centrifugation speeds were considered to represent different centrifugation conditions, i.e., 700 rpm indicated the field water, 2760 rpm indicated the relatively mobile and plant-available water, and 11,000 rpm indicated the permanent wilting point [4]. The solid (soil) and liquid (soil solution) fractions were then collected separately in the glove box under reducing conditions. Finally, the soil solution samples extracted by the above-mentioned CFM were preserved similarly, as carried out for the soil solution samples extracted by the suction method (discussed in Section 2.3) prior to chemical analysis, while the soil samples were stored at 4 °C. A schematic diagram of the CFM is shown in Figure 1.

2.5. Analysis of Soil Solution

Soil solution samples extracted by the suction method and/or CFM were analyzed for their pH and As, Fe(II), and Mn concentrations. The pH was determined using a digital pH meter (LAQUAtwin-pH-33, HORIBA Scientific, Kyoto, Japan). The concentration of total As was measured by ICP-MS (ELAN^® DRC-e, PerkinElmer, Waltham, MA, USA), whereas the concentration of total Mn was quantified by ICP-OES (5110, Agilent Technologies, Santa Clara, CA, USA). The concentration of Fe(II) was determined using a UV spectrophotometer (V-630, JASCO Corporation, Tokyo, Japan). For this, 1 M sodium acetate and 0.2% o-phenanthroline were added during the dilution of the preserved soil solution. To allow o-phenanthroline to fully react with Fe(II), the solution was then kept in an incubator at 30 °C for 30 min. The concentration was then determined from the absorbance at 510 nm.

2.6. Statistical Analysis

Statistical analyses including a correlation coefficient analysis (CCA) were performed using BellCurve for Excel (Version: 3.21), Social Survey Research Information Co., Ltd. (Shinjuku, Tokyo, Japan). As the normal distribution of the sample could not be confirmed, a non-parametric analysis, Kruskal–Wallis’s test, was performed to detect whether there was any statistically significant (p < 0.05) difference among the solute concentrations in the soil solution extracted by the suction method and CFM at different centrifugation speeds.

3. Results

3.1. Soil Properties

The physicochemical properties of the collected soil sample are shown in

Table 1. Physicochemical properties of soil sample.

Parameters	Extraction Methods	Values
Sand	-	67.1%
Silt	-	17.6%
Clay	-	15.3%
pH	H2O	5.27
	KCl	4.01
EC	-	0.04 dS/m
TC	-	15 g/kg
TN	-	1.35 g/kg
C/N	-	11.1
Mg2+	-	0.71 cmolc/kg
Ca2+	-	4.41 cmolc/kg
K+	-	0.23 cmolc/kg
Na+	-	0.13 cmolc/kg
CEC	-	10.4 cmolc/kg
BSP	-	52.7%
Available phosphate	-	92.8 mg P2O5/kg
Feo	Acid ammonium oxalate	2.31 g/kg
Alo		1.68 g/kg
Sio		0.476 g/kg
Fep	Sodium pyrophosphate	1.73 g/kg
Alp		0.846 g/kg
Fed	DCB	3.96 g/kg
Ald		1.19 g/kg
As	HCl	2.03 mg/kg

. The soil shows a relatively high sand content, and the soil texture is sandy clay loam (SCL). The soil was classified as Epiaquepts (soil type) based on the United States Department of Agriculture (USDA) Soil Taxonomy system.

3.2. Changes in Soil pH and Eh during Incubation Experiments

Soil pH was found to increase with time during the incubation experiments, whereas the highest pH (~6.0) was observed in soil treated with just water followed by OM (~5.4), i.e., ethanol and molasses (Figure 2). The lower pH in the soil treated with OM can be partiality explained by the low pH of the OM itself. On the other hand, a decrease in soil Eh (water: +212 to −397 mV; ethanol: +187.5 to −323.5 mV; molasses: +261.5 to −358 mV) was observed with incubation time (day 0 to 18). Noticeably, a sharp decline in soil Eh was observed within the first 3 days of incubation, especially for the samples treated with ethanol and molasses, while a comparatively slower decrease in Eh was observed when just water was added (Figure 2). The decrease in Eh was relatively faster until day 5, although no significant difference was found among the treatments from day 5 to 18.

3.3. Behavior of Solutes (Metals) during Incubation Experiments

With the increasing soil pH and decreasing soil Eh, the solute (As, Fe(II), and Mn) concentrations in the soil solution (extracted by the suction method) increased sharply, regardless of the treatments applied, compared to adding just water (Figure 3). Interestingly, the addition of OM significantly increased the metal concentrations in the soil solution compared to the no-material treatment (water). The results showed that the addition of ethanol increased As, Fe(II), and Mn concentrations by 71, 106, and 118%, respectively, whereas the addition of molasses increased these corresponding concentrations by 43, 83, and 70%, respectively (Figure 3).

The increased metal concentrations with the increasing soil pH and decreasing Eh, especially under OM treatments, indicated that the behavior and chemistry of As, Fe(II), and Mn are closely interrelated and influenced by similar edaphic factors. The CCA also exhibited very strong positive correlations between As, Fe(II), and Mn (r = 0.902–0.951) (

Table 2. Correlation matrix showing the interrelationships between the studied solutes (As, Fe(II), and Mn) in the soil solution.

	As	Fe(II)	Mn
As	-	-	-
Fe(II)	0.914	-	-
Mn	0.902	0.951	-

).

3.4. Effects of Centrifugation Time and Speed on Soil Solution Extraction

For our study, a shorter centrifugation time was crucial to avoid the unintentional oxidation of soil samples, while the extraction of the maximum amount of soil solution was another important consideration. Therefore, a preliminary experiment was performed using random wet soil samples to evaluate whether centrifugation time has any significant effect on the amount of soil solution extracted. The results revealed that centrifugation for more than 1 min resulted in no significant (p > 0.05) increase in the amount of extracted soil solution, which was true for all three centrifugation speeds. While centrifugation for only 1 min extracted 3.045, 10.855, and 18.35 g of soil solution at 700, 2760, and 11,000 rpm, respectively, the centrifugation, when continued for 7 min, extracted a cumulative of only 4.925, 13.895, and 20.3 g of soil solution at those corresponding centrifugation speeds (Figure 4).

Once the preliminary experiment confirmed that a longer centrifugation has no significant effect on the amount of soil solution extracted (Figure 4), a shorter centrifugation time (i.e., 1 min) was used for the soil solution extraction, by CFM, from incubated soil samples at three different centrifugation speeds. The results revealed that a higher centrifugation speed resulted in a higher amount of soil solution extracted (Figure 5), as was also found during the preliminary experiment (Figure 4). Centrifugation at 11,000 rpm for 1 min resulted in a 5- and 1.4-times higher amount of soil solution compared to that obtained by centrifugation for 1 min at 700 and 2760 rpm, respectively (Figure 5).

3.5. Solute Concentrations in CFM-Extracted Soil Solutions

The results exhibited no significant difference in solute (As, Fe(II), and Mn) concentrations when the varying amounts of soil solution extracted by CFM at different centrifugation speeds were measured (Figure 6). The relationship between the solute concentrations and amount of CFM-extracted soil solution was found to be identical to that observed in the soil solution extracted by the suction method (Figure 3), while significantly higher solute concentrations were measured in the OM-treated samples compared to the samples treated with just water (Figure 6). Moreover, Kruskal–Wallis’s test revealed that there was no significant (p > 0.05) difference between the solute concentrations detected in soil solutions extracted by CFM at different centrifugation speeds and the solute concentrations extracted by the suction method (

Table 3. Comparison between the solute concentrations in the CFM-extracted soil solutions and the solute concentrations in the soil solution extracted by the suction method, under different treatments.

Treatments	Extraction Method	Solute Concentrations (mg/L)
		As	Fe(II)	Mn
Water	CFM-700 rpm	0.101 ± 0.001 a	55.369 ± 2.480 a	1.753 ± 0.280 a
	CFM-2760 rpm	0.105 ± 0.016 a	49.012 ± 4.185 a	1.630 ± 0.351 a
	CFM-11,000 rpm	0.113 ± 0.002 a	52.081 ± 2.945 a	1.907 ± 0.045 a
	Suction method	0.142 ± 0.002 a	64.466 ± 1.705 a	2.381 ± 0.110 a
	Test value	χ2 = 4.667	χ2 = 5.500	χ2 = 4.500
	p value	0.198	0.139	0.212
Ethanol	CFM-700 rpm	0.238 ± 0.002 a	134.610 ± 0.000 a	5.258 ± 0.079 a
	CFM-2760 rpm	0.197 ± 0.016 a	118.280 ± 6.355 a	4.239 ± 0.378 a
	CFM-11,000 rpm	0.183 ± 0.021 a	114.773 ± 5.270 a	4.010 ± 0.326 a
	Suction method	0.238 ± 0.000 a	132.966 ± 1.240 a	5.180 ± 0.023 a
	Test value	χ2 = 5.500	χ2 = 6.241	χ2 = 6.167
	p value	0.139	0.101	0.104
Molasses	CFM-700 rpm	0.198 ± 0.000 a	129.843 ± 5.037 a	3.989 ± 0.098 a
	CFM-2760 rpm	0.193 ± 0.001 a	113.183 ± 2.402 a	3.800 ± 0.305 a
	CFM-11,000 rpm	0.199 ± 0.001 a	118.389 ± 3.100 a	4.038 ± 0.274 a
	Suction method	0.207 ± 0.003 a	117.841 ± 3.565 a	4.052 ± 0.140 a
	Test value	χ2 = 6.667	χ2 = 6.167	χ2 = 1.167
	p value	0.083	0.104	0.761

Note: the lowercase letter (a) indicates that there is no significant (p > 0.05) difference between the solute concentrations in soil solutions extracted by the CFM and suction method.

).

4. Discussion

Soil pH and Eh were found to increase and decrease, respectively, with time during the incubation experiments (Figure 2). Similar results were also reported by Kashem and Singh, who studied the effects of OM’s application on soil pH and Eh in metal-contaminated soils under flooded conditions using a growth chamber experiment [22]. Under waterlogged conditions, soils usually have an Eh of <+350 mV [23,24]. Depending on the Eh, soil conditions can also be categorized as moderately reduced soils (+100 to +400 mV), reduced soils (−100 to +100 mV), and highly reduced soils (−100 to −300 mV) [23]. On the other hand, soil pH and Eh are known to be negatively correlated [23,25], as a decrease in Eh is accompanied by an increase in pH as a result of reduction reactions, which consume H⁺ [22,26]. Thus, the increase and decrease in soil pH and Eh, respectively, with time clearly demonstrated that the soil’s reductive condition was maintained and essentially promoted over time during the incubation experiments. Since the application of OM promotes soil reductive conditions by accelerating anaerobic microbial activities [21], the addition of ethanol or molasses promoted the soil’s reductive conditions, as indicated by a sharp decrease in soil Eh for the first 3 days until day 5 (Figure 2), although the OM did not show any significant effect in later days (day 5 to 18).

Such an increase in the concentrations of As, Fe(II), and Mn during the incubation can be explained by two mechanisms occurring under waterlogged reductive conditions: (i) the reduction of oxyhydroxides of Fe and Mn, which results in the release of sorbed As into the liquid phase, and (ii) the reduction of As(V) sorbed on the solid phase (mineral surfaces) to As(III), which thereby remains available in the solution phase [27,28,29]. Moreover, significantly higher solute concentrations in the OM-amended samples supported the idea that the addition of OM substantially enhances the reduction of As, Fe, and Mn under reduction conditions essentially through promoting microbial activity [21,30,31,32].

Under centrifugation for more than 1 min, no significant increase in the volume of the soil solution observed (Figure 4), which implied that centrifugation for 1 min is enough to extract most of the soil solution by CFM when the soil texture is sandy clay loam (Table 1). However, longer centrifugation times are probably needed to collect a sufficient volume of soil solution from clayey soils because of their higher soil matric potential [3]. Nevertheless, the results demonstrated that centrifugation speed, and not centrifugation time, is the predominant factor for extracting the maximum amount of soil solution (Figure 5 and Figure 6), which was found to be in line with the results of Fraters et al. [4]. The major effect of centrifugation speed and the lesser effect of centrifugation time on soil solution extraction can be attributed to the level of compaction, which occurs during centrifugation, with an increase in dry bulk density and a reduction in soil pore size, and is essentially dependent on centrifugation speed [4]. The faster and stronger compaction of soil at a higher centrifugation speed (11,000 rpm) possibly decreased the soil pore size and increased the flow of water [33,34], which thereby resulted in a higher volume of soil solution.

The results demonstrated that there will be no significant difference in solute (As, Fe(II), and Mn) concentrations even though varying amounts of soil solution are extracted by CFM at different centrifugation speeds (700, 2760, and 11,000 rpm). Additionally, the As, Fe(II), and Mn concentrations in the CFM-extracted soil solution showed identical chemical behavior for each tested treatment (Figure 6), which was also in line with the results observed for the solute concentrations in the soil solution extracted by the suction method (Figure 3). The changing of soil reductive conditions to oxidation conditions results in a decrease in the concentrations and availability of As, Fe, and Mn in the soil solution [8,27]. Interestingly, no significant difference was observed between the solute (As, Fe(II), and Mn) concentrations in the CFM-extracted soil solution and the corresponding solute concentrations in the soil solution extracted by the suction method (Table 3), confirming that the reductive conditions were maintained during the extraction of soil solution by CFM. All these results suggested that a CFM operating at a higher centrifugation speed (11,000 rpm) could potentially be employed as an effective method for the efficient separation of solid and liquid phases of WRS within a short time period while maintaining its reductive conditions, maximizing the amount of soil solution extracted and not affecting its solute concentrations.

5. Conclusions

Although various established methods are available for the separation of the solid–liquid phases of upland soils, there is no established method in the literature for effectively separating the solid and liquid phases of WRS while maintaining its reductive conditions. This study verified that the solid and liquid phases of WRS can be efficiently separated by CFM while maintaining its reductive conditions. The results demonstrated that a higher centrifugation speed (11,000 rpm), representing the permanent wilting point, would be the most effective for the efficient separation of solid–liquid phases while extracting most of the soil solution within a short time (i.e., 1 min). Moreover, our findings on the solute concentrations in the extracted soil solution confirmed that the application of CFM would greatly help us to quickly separate and study the chemistry of solid and liquid phases of WRS. However, further research is needed to verify the effectiveness of CFM on other soil types, as well as to widely apply this simple solid–liquid separation method for a better understanding of the chemical behavior and dynamics of the solutes in soil solutions under reductive conditions.