Fixation and defixation of ammonium in soils: a review

Abstract

Fixed NH ⁺₄ (NH ⁺_{4 f}) and fixation and defixation of NH ⁺₄ in soils have been the subject of a number of investigations with conflicting results. The results vary because of differences in methodology, soil type, mineralogical composition, and agro-climatic conditions. Most investigators have determined NH ⁺_{4 f} using strong oxidizing agents (KOBr or KOH) to remove organic N and the remaining NH ⁺_{4 f} does not necessarily reflect the fraction that is truly available to plants. The content of native NH ⁺_{4 f} in different soils is related to parent material, texture, clay content, clay mineral composition, potassium status of the soil and K saturation of the interlayers of 2:1 clay minerals, and moisture conditions. Evaluation of the literature shows that the NH ⁺_{4 f}-N content amounts to 10–90 mg kg⁻¹ in coarse-textured soils (e.g., diluvial sand, red sandstone, granite), 60–270 mg kg⁻¹ in medium-textured soils (loess, marsh, alluvial sediment, basalt) and 90–460 mg kg⁻¹ in fine-textured soils (limestone, clay stone). Variable results on plant availability of NH ⁺_{4 f} are mainly due to the fact that some investigators distinguished between native and recently fixed NH ⁺₄ while others did not. Recently fixed NH ⁺₄ is available to plants to a greater degree than the native NH ⁺_{4 f}, and soil microflora play an important role in the defixation process. The temporal changes in the content of recently fixed NH ⁺₄ suggest that it is actively involved in N dynamics during a crop growth season. The amounts of NH ⁺₄ defixed during a growing season varied greatly within the groups of silty (20–200 kg NH ⁺₄ -N ha⁻¹ 30 cm⁻¹) as well as clayey (40–188 kg NH ⁺₄ -N ha⁻¹ 30 cm⁻¹) soils. The pool of recently fixed NH ⁺₄ may therefore be considered in fertilizer management programs for increasing N use efficiency and reducing N losses from soils.

Introduction

Some soils are able to bind ammonium (NH ⁺₄ ) and potassium (K⁺) in such a manner that these cannot be easily replaced by other cations. McBeth (1917) was the first who observed that NH ⁺₄ added to soils could not be completely recovered by alkaline distillation or by extraction with 10% hydrochloric acid. McBeth (1917) defined the unrecovered portion of the added NH ⁺₄ as “fixed NH ⁺₄ ” (henceafter refered as NH ⁺_{4 f}). The fixation of NH ⁺₄ is defined as “the adsorption or absorption of ammonium ions by the mineral or organic fraction of the soil in a manner that they are relatively unexchangeable by the usual methods of cation exchange” (Osborne 1976a; SSSA 1984). Osborne (1976a) suggested “intercalary NH ⁺₄ ” and Mengel and Scherer (1981) “non-exchangeable NH ⁺₄ ” to distinguish between the NH ⁺₄ recovered from clay minerals by digestion of the soil with 5 N HF:1 N HCl and NH ⁺₄ which is bound organically or as NH₄-phosphate (Frye and Hutcheson 1981). While earlier investigations concluded that only a very small amount of NH ⁺_{4 f} is available to microorganisms and plants (Allison et al. 1951, 1953b; Axley and Legg 1960; Lutz 1966), studies in the last three decades suggest that NH ⁺_{4 f} may be released and used by crops (Kudeyarov 1981; Nommik 1981; Mengel and Scherer 1981; 1986; Nommik and Vahtras 1982; Preston 1982; Scherer 1984, 1987, 1993; Lu et al. 2010).

Ammonium fixation and release can play a crucial role for the efficiency of fertilizer N (Scherer and Mengel 1986; Dou and Steffens 1995; Steffens and Sparks 1999; Juang et al. 2001) as it impacts the indigenous soil N supply towards crop N uptake. Nitrogen contributions from soil including defixation of NH ⁺_{4 f} in a given year/season can greatly alter recovery efficiency of applied N because there occurs a large fertilizer N substitution of soil N. With current concerns about the environment and the need to produce more food with less fertilizer N input there is an increased focus for quantifying soil borne N supply. The need to boost agricultural production worldwide is stimulating N fertilizer consumption. In Asia and the Americas, ammonium-based N fertilizers are increasingly used (Prud’homme 2005) and their recovery efficiency by the plants, inter alia, could be influenced by fixation and release of NH ⁺₄ . In soils with high NH ⁺₄ fixation capacity a part of the NH ⁺₄ supplied through NH ⁺₄ -forming or NH ⁺₄ -containing fertilizers may be bound in clay mineral interlayers. Increasing NH ⁺₄ fixation can be a way in building up an available N pool in soils to optimize crop recovery and minimize N losses into the environment (Liu et al. 2008) as the NH ⁺₄ ions after penetration into the clay mineral interlayers are excluded from nitrification (Guo et al. 1983) and are thus protected against leaching. The NH ⁺_{4 f} pool can thus function as a kind of buffer that could influence N losses from soils and mineral N availability to crops.

Some reviews on the fixation and availability of fixed NH ⁺₄ have been published in the past (Rodrigues 1954; Nommik 1957; Bremner 1959; Nommik and Vahtras 1982; Scherer 1993). These reviews mainly focused on the chemistry and factors influencing NH ⁺₄ fixation in soils and gave relatively less attention to the seasonal dynamics and plant availability of NH ⁺_{4 f}. While the chemistry of NH ⁺₄ fixation is fairly well understood (Baethgen and Alley 1987; Nommik and Vahtras 1982), there are conflicting reports on the availability of NH ⁺_{4 f}. This is assumed to be due to differences in methodology, soil type, mineralogical composition, and agro-climatic conditions. It is, therefore, important that the existing information may be integrated with underlying variants for better understanding the role of NH ⁺₄ fixation and release on N dynamics in the soil–plant system.

The present paper gives an overview of the current knowledge on NH ⁺_{4 f} and ammonium fixation and defixation, particularly on pool sizes in different soils with particular emphasis on new aspects on availability of recently fixed NH ⁺₄ to crops and microflora and seasonal changes of NH ⁺_{4 f} pools under plant cover. The information is important not only for fertilizer management but also on the background of increasing importance of NH ⁺₄ -based fertilization, namely urea, and an intensive debate on NH ⁺₄ -based fertilization (cultan fertilization) in plant nutrition (Sommer and Scherer 2009). The paper includes literature of the last few decades, and the soils were grouped according to parent material and texture.

Methods for the determination of NH ⁺_{4 f}

Several methods have been proposed for determining NH ⁺_{4 f} in soil. Most of the methods involve pretreatment of the samples to remove organic matter and exchangeable NH ⁺₄ and dissolution/extraction of the residual sample for release of NH ⁺_{4 f}. Barshad (1951) suggested distilling NH ⁺₄ -treated soil with NaOH and KOH to remove exchangeable NH ⁺₄ . This method resulted in an incomplete recovery in some cases probably because of blocking effect of K⁺ or interference by soil organic matter. In the method by Mogilevkina (1964), soil organic matter is removed by dry combustion at 400°C during time periods between 24 and 72 h, depending on the organic matter content, prior removal of the fixed NH ⁺₄ using sulfuric acid (Kjeldahl procedure). Most of the other methods use HF for the extraction of NH ⁺_{4 f} (Rodrigues 1954; Dhariwal and Stevenson 1958; Bremner 1959). The method proposed by Silva and Bremner (1966) is the most commonly used and involves treating the soil with alkaline KOBr solution (to remove organic compounds), washing the residue with 0.5 M KCl and shaking with 5 N HF:1 N HCl for 24 h. The 24 h time for HF–HCl treatment could be reduced to 30 min by performing the treatment at 100°C. The NH ⁺₄ released is determined by steam distillation of the soil–acid mixture after adding KOH. The pretreatment used in this method to remove organic compounds is considered superior to those used in previous methods for estimating fixed ammonium because it effects an almost entire removal of organic soil N without the risk of fixation of NH ⁺₄ by soil minerals. Zhang and Scherer (1998) suggested replacing the boiling of the mixture of soil and KOBr on a hot plate with keeping the soil–KOBr mixture for 10 min either in a water bath at 95°C or heating in a microwave oven (1,150 W) at 50% of full power for 10 min. Marzadori et al. (1994) boiled the soil–KOBr mixture in a microwave digestion system for 5 min at 90% and 2 min at 80% of the maximum oven energy (600 W). Antisari et al. (1987) used H₂O₂ along with pyrophosphate and NaCl (H₂O₂-NaCl) for the destruction of organic matter and the reproducibility of their method was better than that of the Silva and Bremner (1966) method. The method of Paramasivam and Breitenbeck (2000) excludes the alkaline pretreatment that removes labile organic compounds. Cox et al. (1996) suggested the use of sodium tetraphenylboron (NaBPh₄) that facilitates the release of NH ⁺_{4 f} as NH₄BPh₄. However, compared with the method of Silva and Bremner (1966), only 71% of NH ⁺_{4 f} were extracted. Although the method may extract only a part of the NH ⁺_{4 f}, it may be a better estimate of potentially plant-available NH ⁺₄ (Cox et al. 1996).

Comparison of some of the methods showed that these give widely different values for a given amount of NH ⁺_{4 f}. For example, the method of Mogilevkina (1964) could recover only a quarter while the rest was lost during ignition of the soil sample to remove organic matter (Bremner et al. 1967). Similarly, Moyano and Gallardo (1988) also obtained significantly lower values with the method of Mogilevkina (1964) as compared to that of Silva and Bremner (1966). The method of Silva and Bremner (1966) gave higher values than the Dhariwal and Stevenson (1958), Bremner (1959), and Mogilevkina (1964) methods and lower values than the Rodrigues (1954) and Schachtschabel (1960) methods (Silva and Bremner 1966; Osborne 1976a).

Dixit and Mir (1987) compared the methods proposed by Rodrigues (1954), the modified Rodrigues method (Bremner et al. 1967) and those by Dhariwal and Stevenson (1958), Bremner (1959), and Silva and Bremner (1966) on ten soils varying in clay content from 12% to 31%. The highest values were obtained in the method of Rodrigues and the lowest by the method of Bremner. Similar results were reported by Opuwaribo and Odu (1974). Antisari and Sequi (1988) compared results after the application of three methods (Silva and Bremner’s method, HF–HCl treatment in a microwave system, and CHN analyzer). The microwave system and CHN analyzer produced similar results and the values were generally 1.5 to 2.7 times higher than those obtained with Silva and Bremner’s treatment indicating incomplete recovery by the latter method.

Mechanism of NH ⁺₄ fixation

Ammonium fixation is greatest in 2:1 type clay minerals such as illite, vermiculite, and montmorillonite. Clay minerals possess negative charges balanced by cations, for example, NH ⁺₄ or K⁺. The physics of NH ⁺₄ is closely related to that of K⁺ because both ions have similar ionic radii and low hydration energy. The complete concept with all its further aspects has been nicely put down in detail by Nommik (1965). For both NH ⁺₄ and K⁺, the same mechanism is responsible for fixation and both fit exactly into the ditrigonal holes in the basal oxygen plane of 2:1 clay minerals. The penetration of both cations into the clay mineral interlayers causes the clay layers to collapse to 1 nm, and NH ⁺₄ and K⁺ ions are trapped between silicate sheets and largely withdrawn from exchange reactions (Nommik 1965). Therefore, both cations held in the interlayers of collapsed 2:1 clay minerals are said to be “fixed” and the term “fixed NH ⁺₄ ” was formerly used. The electrostatic energy between NH ⁺₄ (or K⁺) and the negative charges in the crystal sheets is greater than the hydration energy of ammonium. The NH ⁺₄ ion readily sheds its hydration water shell and enters the lattice void, where fixation occurs (Kittrick 1966).

Contents of NH ⁺_{4 f} in soils and influencing factors

Contents of NH ⁺_{4 f}

Contents of NH ⁺_{4 f}-N in the plough layer of arable soils cover a wide range (Table 1). Parent material has a major influence on the NH ⁺_{4 f}-N pool in soils developed from different parent materials, which increases in the order sand (diluvial sand and red sandstone) < basalt ≈ granite < loess ≈ ground moraine < alluvial sediment < limestone < marsh sediment. The high discrepancy in the NH ⁺_{4 f}-N content of soils that developed from clay stone (Mohammed 1979 vs. Zhang et al. 2007) may be mainly due to different clay mineralogy (note that the clay contents given by Zhang et al. (2007) only represent the fraction <1 μm). In summary, the NH ⁺_{4 f}-N content amounts to 10–90 mg kg⁻¹ in coarse-textured soils (e.g. diluvial sand, red sandstone, granite), 60–270 mg kg⁻¹ in medium-textured soils (loess, marsh, alluvial sediment, basalt), and 90–460 mg kg⁻¹ in fine-textured soils (limestone, clay stone).

Table 1 Contents of fixed NH ⁺₄ in soils derived from different parent materials (Source: Nieder and Benbi 2008, extended)

At the regional scale, the contents of NH ⁺_{4 f}-N also vary considerably. For example, the content of NH ⁺_{4 f}-N of different soils in Austria ranged between 45 and 190 mg kg⁻¹ (Schiller and Wallicord 1964), of Spanish soils between 180 and 490 mg kg⁻¹ (Moyano and Gallardo 1988), of soils in Turkey between 60 and 230 mg kg⁻¹ (Elmaci et al. 2002), of soils in Queensland (Australia) between 6 and 107 mg kg⁻¹ (Osborne 1976b), of soils of the former USSR between 40 and 490 mg kg⁻¹ (Skonde et al. 1974), of soils in China between 35 and 573 mg kg⁻¹ (Qi-Xiao et al. 1995), of soils in Nigeria between 8 and 98 mg kg⁻¹ (Opuwaribo and Odu 1974), and of soils in southern Ontario (Canada) between 57 and 367 mg kg⁻¹ non-exchangeable NH₄-N (Doram and Evans 1983). The content in soils in Kentucky (USA) was as high as 365 mg kg⁻¹ (Sparks et al. 1979). Investigations on top soils from Antarctic also showed that NH ⁺_{4 f}-N occurs in amounts (0 to 322 mg kg⁻¹) similar to elsewhere in the world (Greenfield 1991).

The percentage of total N present as NH ⁺_{4 f}-N in main soil groups of Israel ranged between 1.8% and 78.6%, with most of the values falling between 2% and 25%. Highest values were obtained for the deepest horizons (Feigin and Yaalon 1974), where it ranged from 14% to 78% in British Caribbean soils (Rodrigues 1954), 3% to 44% in southern Ontario soils (Doram and Evans 1983), 16% to 59% in Vertisols and 13% to 31% in Cambisols in India (Sahrawat 1995), 21% in Italian Fluvisols (Benedetti et al. 1996), and up to 85% in subsoils in the USA (Smith et al. 1994).

In a soil profile, the NH ⁺_{4 f}-N content as percent of the total N (inorganic plus organic) generally increases with soil depth (Black and Waring 1972; Doram and Evans 1983; Smith et al. 1994; Sahrawat 1995; Zhang et al. 2003) due to decreasing soil organic matter content. Fixation of NH ⁺₄ in soil organic matter is negligible (Kowalenko and Cameron 1976). Contents of NH ⁺_{4 f}-N in 24 soils from Queensland (Australia) averaged 4% of total N in surface soil and 6.4% for subsoils (Black and Waring 1972). The magnitude of increase with depth was much higher in soils from southwestern Saskatchewan where it ranged from 7% of the total N in the surface soil to 58% in 120 cm depth (Hinman 1964). In Spanish soils, it ranged from 21% to 33% in the surface layers and 30% to 83% in subsoils with illite as the dominant clay mineral (Moyano and Gallardo 1988).

Yaalon and Feigin (1970) found that illite contains approximately 600 mg N kg⁻¹ illite, whereas montmorillonite contains small and kaolinite clays only negligible amounts. In mixed soil clays, the NH ⁺_{4 f}-N level is determined by the quantity of illite (Feigin and Yaalon 1974) and illite plus vermiculite (Sparks et al. 1979; Doram and Evans 1983), respectively. In contrast to the above results, Bajwa (1985) in a comparative study on fixation of NH ⁺₄ and K⁺ in wetland rice soils found that in five soil clays, montmorillonitic clay fixed the maximum of the added NH ⁺₄ (98%) followed by vermiculite (88%) and least (34%) in clay containing hydrous mica, chlorite, and halloysite. On the other hand, montmorillonite, contrary to the other clay minerals, had no capacity for fixing K⁺. The overall capacity for fixation of monovalent cations (K⁺ plus NH ⁺₄ ) may therefore be greater for vermiculite compared to montmorillonite.

Influencing factors

The content of NH ⁺_{4 f} and the capacity of soils to fix NH ⁺₄ is related to parent material (Table 1), texture (Baethgen and Alley 1987), clay content (Opuwaribo and Odu 1974; Sowden et al. 1978; Moyano and Gallardo 1988; Juang 1990), clay mineral composition (Feigin and Yaalon 1974; Sparks et al. 1979; Doram and Evans 1983; Niederbudde 1983), the K concentration in soil solution, the degree of K saturation of the exchange complex of the soil colloids and K saturation of the interlayers of 2:1 clay minerals (Hinman 1966; Doram and Evans 1983), and soil moisture conditions (Black and Waring 1972). Allison et al. (1953a) showed that vermiculite and illite have the greatest capacity to fix NH ⁺₄ , while montmorillonite fixed less NH ⁺_{4 f} and held it less tenaciously. Said (1973) reported that soils from Sudan with montmorillonite as the dominant clay mineral contained only small amounts of NH ⁺_{4 f}-N ranging between 30 and 60 kg N kg⁻¹ soil. As compared to montmorillonite, beidelite is a high fixing smectite, due to the isomorphic substitution in the tetrahedral layer (Feigenbaum et al. 1994). Kaolinites, belonging to the 1:1 type of clay minerals, are not able to bind NH ⁺₄ ions in their interlayers because hydrogen bonds which join the interlayers allow only very little dilatation of the narrow interlayer space (Mela Mela 1962).

In addition to clay minerals, the silt fraction has also been reported to bind NH ⁺₄ -N in a non-exchangeable form. For example, in loess soils of central Europe, about only 65% of the NH ⁺_{4 f}-N is interlayer ammonium (Niederbudde and Friedrich 1984), while the rest is found in the silt fraction. The content of NH ⁺_{4 f}-N in the clay and silt ranged from 255 to 430 mg N kg⁻¹ and from 72 to 166 mg N kg⁻¹, respectively (Jensen et al. 1989). In a Canadian soil with 37% clay, Kowalenko and Ross (1980) found similar amounts of NH ⁺₄ -N for the clay and the silt fraction. In different primary minerals, the content of NH ⁺_{4 f}-N has been reported to vary between nil in quartz to 266 mg kg⁻¹ in biotite (Wlotzka 1961).

Some studies have demonstrated that the NH ⁺₄ fixation capacity strongly depended on the degree of K saturation of the interlayers of the 2:1 clay minerals. If the K content of a soil is high, it can be expected that the interlayer space will also be saturated with potassium (Scherer 1982) and to a smaller extend with NH ⁺₄ ions (Petersburgsky and Smirnov 1966). As a result of competition for fixation sites, the presence of NH ⁺₄ or K⁺ may alter both fixation and release of these cations. In several studies, addition of K⁺ prior to NH ⁺₄ depressed NH ⁺₄ fixation (Stanford and Pierre 1947; Nommik and Vahtras 1982), and addition of NH ⁺₄ prior to or at the same time as K⁺ reduced K⁺ fixation (Aquaye and MacLean 1966; Bartlett and Simpson 1967). Contrarily, Drury et al. (1989) found that K⁺ pre-addition did not block subsequent NH ⁺₄ fixation and the presence of K⁺ induced greater NH ⁺₄ fixation (Chen et al. 1989). This is in agreement with the study by Bajwa (1985) in that the sequence in which NH ⁺₄ and K⁺ were applied did not appear to affect the relative amounts that were fixed.

Application of NH ⁺₄ -containing fertilizers may result in fixation of the added NH ⁺₄ . However, the amount fixed is not directly proportional to the amount of fertilizer N added (Liang and MacKenzie 1994). Fixation is usually fast and occurs within the first few hours after fertilizer application. The fixation rate is controlled mainly by ion diffusion and declines with time until the equilibrium point is approached (Nommik 1965). In experiments under laboratory conditions, Kowalenko and Cameron (1976) found that 50% of added NH ⁺₄ was fixed within a short period of time, while in a field experiment, Kowalenko (1978) found that 59% of 152 kg N ha⁻¹ added as (NH₄)₂SO₄ was fixed within a few minutes after addition. In some soils of eastern Canada, more than 40% of the added NH ⁺₄ was fixed in less than 2 h and further fixation was negligible (Sowden et al. 1978). According to Chantigny et al. (2004), the highest fixation of NH ⁺₄ by clay minerals occurred within the first day and was greater in the clay soil (34% of applied ¹⁵NH₄-N) than in the sandy soil (11%). In studies with ¹⁵N-labeled NH ⁺₄ 18% to 23% of the applied ¹⁵NH ⁺₄ was fixed after a 15-day incubation in soils with high contents of vermiculite (Drury et al. 1989).

The amount of added NH ⁺₄ fixed depends on the NH ⁺₄ -fixing capacity of the soil. The estimation of the NH ⁺₄ fixation capacity of a soil involves treatment of the soil with excessive amounts of concentrated NH ⁺₄ solution in order to achieve complete saturation of the fixing minerals with NH ⁺₄ (Nommik and Vahtras 1982) before estimation of the fixed NH ⁺₄ by one of the methods described above. In a Gleysol with high NH ⁺₄ -fixing capacity, Fischer et al. (1981) found that 47% of added ¹⁵N were fixed, whereas in a Histosol with very low NH ⁺₄ fixation capacity, only 7% of the added NH ⁺₄ entered the interlayers of the clay minerals. Rider et al. (2005) focusing on the occurrence of NH ⁺₄ fixation on a decomposed granitic substrate showed that the fixation capacities of these sandy saprolites are significant. At field loading rates equivalent to less than 300 kg NH ⁺₄ -N ha⁻¹ 36% to 42% of the added N became unavailable to plants due to interlayer collapse and fixation. Ammonium fixation did not vary significantly in relation to substrate weathering class in these samples.

According to Allison et al. (1953a), soil moisture can reduce NH ⁺₄ fixation as clay minerals are expanded under wet conditions. However, in dry soils, the interlayer space is reduced and NH ⁺₄ fixation also increases. According to Osborne (1976b), NH ⁺₄ fixation was reduced by 25% in a clay soil moistened to 60% of the maximum water holding capacity as compared with the dry soil. Also, Gouveia and Eudoxie (2007) found a lower NH ⁺₄ fixation in wet soils. However, investigations in flooded rice soils yielded contradictory results. Some authors (Keerthisinghe et al. 1984; Chen et al. 1987a) observed that NH ⁺₄ fixation in paddy soils was significantly increased as a consequence of flooding, while others (Williams et al. 1968; Chen et al. 1987b) stated that NH ⁺₄ fixation was less pronounced under flooded compared to non-flooded conditions. Schneiders and Scherer (1996) observed an increasing concentration of NH ⁺_{4 f}-N after flooding, which was closely related to the declining redox potential (E _h). Therefore, E _h may have an important effect on the fate of NH ⁺₄ in paddy soils (Zhang and Scherer 1999). At a low E _h, the octahedral Fe³⁺ in the clay minerals is assumed to be reduced, resulting in a higher negative charge of the unit cell and therefore a higher Coulombic attraction between the interlayer cations and the silicate layers (Stucki et al. 1984). This increasing NH ⁺₄ fixation can be an option for building an available N pool in paddy soils to minimize N losses into the environment (Liu et al. 2008). A further prerequisite for the pronounced NH ⁺₄ fixation in flooded soils is the microbial reduction of Fe³⁺, followed by the dissolution of Fe oxides coated on the surface of clay minerals at a low E _h, promoting the diffusion of NH ⁺₄ ions into the interlayers of the clay minerals (Scherer and Zhang 1999). Because of the reversible oxidation and reduction of Fe oxides in paddy soils, this mechanism may be of special importance for fixation of NH ⁺₄ (Zhang and Scherer 2000; Scherer and Zhang 2002). Figure 1 shows the relationship between the contents of NH ⁺_{4 f}-N and clay-bound Fe²⁺ in paddy soils and clay minerals.

Fig. 1

Relationship between NH ⁺_{4 f}-N and Fe²⁺ in soils and clay minerals (source: Scherer and Zhang 2002)

Availability of fixed NH ⁺₄ to plants and microflora

Fixation and release of NH ⁺_{4 f}-N is dependent upon chemical equilibria between the amounts of NH ⁺_{4 f}, exchangeable NH ⁺₄ and NH ⁺₄ in soil solution (Nommik and Vahtras 1982),

$$ \left( {{\hbox{N}}{{\hbox{H}}_4}{{^{+} }_{\rm{s}}}} \right) \leftrightarrow \left( {{\hbox{N}}{{\hbox{H}}_4}{{^{+} }_{\rm{e}}}} \right) \leftrightarrow \left( {{\hbox{N}}{{\hbox{H}}_4}{{^{+} }_{\rm{f}}}} \right) $$

where NH ⁺_{4 s} denotes ammonium in the soil solution and NH ⁺_{4 e} and NH ⁺_{4 f} are exchangeable and fixed ions, respectively. However, results from several studies indicate that there is no equilibrium between exchangeable and fixed NH ⁺₄ . Because NH ⁺₄ and K⁺ compete for the same binding sites, Nommik (1957) suggested that the sum of NH ⁺₄ and K⁺ rather than NH ⁺₄ alone should be introduced into the equilibrium equation. Steffens and Sparks (1997) described the kinetics of NH ⁺_{4 f}-N release from soil using the Elovich model. This model is empirical, and the process is assumed to be diffusion controlled. With decreasing NH ⁺₄ concentration in soil solution, NH ⁺₄ ions diffuse from clay mineral interlayers. Therefore, factors such as fertilizer N application, plant cover, soil organic matter, microflora, clay content, and clay mineral composition that affect concentration of NH ⁺₄ in soil solution may promote either release or fixation of NH ⁺₄ . Because of methodological limitations (such as proper separation of recently fixed NH ⁺₄ from native NH ⁺_{4 f}) and the involvement of several factors and processes influencing soil N dynamics, it is still difficult to describe the dynamics of fixed NH ⁺₄ especially under field conditions.

Plant availability

Investigations focusing on the availability of NH ⁺_{4 f}-N for crops have been reported with conflicting results. While early investigations of Legg and Allison (1959) and Black and Waring (1972) found that this N fraction plays a minor role for the N nutrition of plants, Norman and Gilmour (1987) estimated that the amounts of fertilizer-derived NH ⁺_{4 f}-N available to ryegrass ranged from 35% to 72%. As percentage of total NH ⁺_{4 f}-N, the release of this fraction ranged from 4% to 25% in different soils (Osborne 1976b; Smith et al. 1994; Steffens and Sparks 1997). According to Smith et al. (1994), only 8% of NH ⁺_{4 f}-N were released on average from the surface layers of a number of US soils. In a Luvisol derived from loess, up to 250 kg NH ⁺_{4 f}-N ha⁻¹ were released during a growing season of annual crops (oats and winter wheat; Van Praag et al. 1980). According to Baethgen and Alley (1987), NH ⁺_{4 f}-N contributed significantly to N taken up by wheat grown in the greenhouse.

Various plant species influence the NH ⁺_{4 f} pool through different mechanisms. Plants can take up soluble or exchangeable NH ⁺₄ in the vicinity of NH ⁺₄ -fixing clays and thus promote diffusion of ions out of the interlayers. Plant roots can also affect the NH ⁺_{4 f} pool indirectly by releasing exudates that promote the activity of the soil microflora and microbial N uptake (Marschner 1995). The magnitude of NH ⁺₄ -N release strongly depends on the length of crop growth period and plant density (Saha and Mukhopadhyay 1986).

Most of the differences in NH ⁺₄ -N release can probably be attributed to both the variable pool sizes of “native fixed NH ⁺₄ ” and “recently fixed NH ⁺₄ ” and the lack of methods available to separate these fractions adequately. The recently fixed NH ⁺₄ is mainly derived from mineral N fertilizers (Chen et al. 1989; Smith et al. 1994), but it may also originate from mineralization of soil organic matter (Nieder et al. 1995b). Dou and Steffens (1995) found that 90% to 95% of recently fixed ¹⁵NH ⁺₄ was released during a 14-week period in the soil planted with perennial ryegrass (Lolium perenne L.) under greenhouse conditions. Under field conditions, 66% of the recently fixed NH ⁺₄ -N was released in the first 86 days after fixation and the remaining was strongly fixed over the next 426 days (Kowalenko 1978). Apparently, recently fixed NH ⁺₄ resulting from fertilizer application is more available to plants than native NH ⁺_{4 f}, which is more tightly held (Black and Waring 1972; Kudeyarov 1981; Mengel and Scherer 1981; Keerthisinghe et al. 1984). Probably native NH ⁺_{4 f}-N is trapped in the center of the interlayers to a higher degree, while recently fixed NH ⁺₄ ions are largely retained in the peripheral zone of the interlayers (Nommik and Vahtras 1982). Differences in clay mineralogy and K saturation of the minerals also influence the release of NH ⁺_{4 f}-N. Mengel et al. (1990) found that only soils containing vermiculite and low contents of exchangeable K⁺ released significantly higher amounts of NH ⁺_{4 f} as compared to soils with no vermiculite and high percentage of K⁺ saturation of the clay minerals. In a pot experiment of Scherer (1985) with ryegrass, where NH ⁺₄ was added before and after the application of K⁺, respectively or simultaneously, the content of NH ⁺_{4 f} was the highest, when NH ⁺₄ was applied before K and lowest when K was applied first (Table 2).

Table 2 Influence of the application time of K⁺ on the content of NH ⁺_{4 f}-N and exchangeable K⁺ (source: Scherer 1985)

Scherer (1987) further reported that in field plots heavily dressed with K⁺, no release of NH ⁺_{4 f}-N was found, whereas in plots with low K⁺ application, considerable amounts of NH ⁺₄ were released. In analogy to these findings, adsorption/desorption experiments (isotherm studies in soils containing vermiculite and illite) in a binary (Ca²⁺/NH ⁺₄ ) and a ternary (Ca²⁺/K⁺/NH ⁺₄ ) cation system (Lumbanraja and Evangelou 1994) revealed that NH ⁺₄ desorption decreased in the presence of solution K⁺ (ternary) relative to that in the absence of K⁺ (binary). Fixation of K⁺ in the presence of solution NH ⁺₄ was suppressed as compared to that in the absence of NH ⁺₄ . Contrarily, NH ⁺₄ fixation was enhanced in the presence of solution K⁺ relative to that in the absence of K⁺. Thus, contrary to common belief, the two ions do not behave as true analogs with respect to fixation reactions.

The mechanisms regulating the release and subsequent plant uptake are still not completely understood. According to Mengel et al. (1990) and Scherer and Ahrens (1994; 1996), plant roots deplete the NH ⁺₄ concentration of the soil solution in the rhizosphere and, therefore, promote the release of NH ⁺_{4 f} (Fig. 2).

Fig. 2

Depletion profile of NH ⁺_{4 f}-N in the rhizosphere of oilseed rape in a soil with high amount of expandable clay minerals (57% clay and 17% of smectites and vermiculite in the clay fraction; source: Scherer and Ahrens 1996)

Further, the relative amounts of NH ⁺₄ and K⁺ in soil solution govern the rate of release of NH ⁺₄ from interlayer positions (Welch and Scott 1960). Under field conditions, continuous uptake of NH ⁺₄ and K⁺ ions by roots may reduce concentrations of both ions and therefore, diminish the blocking effect of K⁺ on the release of NH ⁺₄ .

Availability to microflora

Early work concerning the influence of the soil microflora on the mobilization of NH ⁺_{4 f} suggested only a minor role for the microflora (Allison et al. 1953b). While only between 13% and 28% of the NH ⁺_{4 f} was oxidized, up to almost 80% of the exchangeable NH ⁺₄ was nitrified (Bower 1950). However, according to later studies heterotrophic microorganisms can rapidly assimilate NH ⁺₄ from the non-exchangeable pool (Nommik and Vahtras 1982; Drury and Beauchamp 1991) and therefore favor the release of NH ⁺₄ ions from the clay minerals (Jensen et al. 1989). In incubation experiments involving ¹⁵N-labeled NH ⁺_{4 f}, the addition of an easily available C substrate favored the mobilization of this N fraction by heterotrophic microorganisms (Breitenbeck and Paramasivam 1995), so that between 64% and 96% of non-exchangeable ¹⁵NH ⁺₄ was released. Different plant species produce root exudates with different quantity and quality which can differently affect activity and composition of the soil microflora (Nannipieri et al. 1999). Moreover, the release of NH ⁺_{4 f} can be promoted by nitrification which causes changes in NH ⁺₄ equilibria (Green et al. 1994). Fumigation to inhibit soil biological activity as well as the application of a nitrification inhibitor (nitrapyrin) hampered the release of NH ⁺_{4 f} (Aulakh and Rennie 1984). According to Tang et al. (2008), the stimulated assimilation of ammonium after glucose addition creates a steep concentration gradient between the NH ⁺₄ concentration of the soil solution and the NH ⁺_{4 f} and thus promoting the release of NH ⁺₄ from the clay mineral interlayers. The results also show that K added simultaneously with the carbon source impeded the release of NH ⁺₄ ions, which may be due to the blocking effect of K⁺ ions.

Results from field experiments during the cereal growing season (Nieder et al. 1995a, b, 1996) suggest that the dynamic microflora, due to its influence on the equilibria between the amounts of different N fractions, may exert a great influence on the dynamics of NH ⁺_{4 f}. The minima of mineral N contents (due to plant and microbial N uptake) and correspondingly, of NH ⁺_{4 f} contents, occurred during phases of increasing microbial biomass under a C-supplying plant cover, proving that the heterotrophic microflora, as a consequence of the increased microbial N uptake, favor the release of NH ⁺_{4 f} (Fig. 3).

Fig. 3

Dynamics of NH ⁺_{4 f}-N, microbial biomass N, and mineral N (NH ⁺₄ -N plus NO ⁻₃ -N) in a Luvisol A_p horizon under winter wheat (source: compiled from Nieder et al. 1995a)

Seasonal dynamics of NH ⁺_{4 f}

The pattern of seasonal changes in the contents of NH ⁺_{4 f} in arable soils is well-known (Kowalenko and Cameron 1976; Sowden 1976; Kowalenko and Ross 1980; Li et al. 1990a, b; Drury and Beauchamp 1991; Green et al. 1994). Mengel and Scherer (1981), investigating the dynamics of this N fraction in a Fluvisol during the growing season, found that the content of NH ⁺_{4 f} declined in the top 60 cm from February to May (Fig. 4).

Fig. 4

Release of NH ⁺_{4 f}-N in a Fluvisol during the growing season of spring oats (source: Mengel and Scherer 1981)

In the deeper soil layer (60–90 cm depth), depletion was observed from May to July, which was in accordance with the root growth of spring oats. At the end of the growing season, the clay mineral interlayers were refilled and almost the same content was attained as in spring. Li et al. (1990a, b) have shown a significant decrease of NH ⁺_{4 f} in upper soil layers in March and in deeper soil layers during April. In field experiments on loess soils, Nieder et al. (1996) observed a significant correlation between the time course of NH ⁺_{4 f} and mineral N in soil solution. The refixation of NH ⁺₄ in autumn may be mainly due to increased mineralization after the harvest of annual crops. Table 3 reflects this pattern (ranges of NH ⁺₄ fixation in spring, NH ⁺₄ release in summer and NH ⁺₄ refixation in autumn) with results (according to pool size or ¹⁵N method, respectively) drawn from different studies, grouped according to upland (silty and clayey) and paddy soils.

Table 3 Fixation, release and refixation of NH ⁺₄ -N in the 30 cm A_p horizon of arable loess soils during the growth period of annual crops as estimated by the temporal variation in pool size and the ¹⁵ N method

In most of studies presented in Table 3, the NH ⁺₄ -N fixation in spring, occurring probably as a consequence of (mineral) fertilizer application at the beginning of the growth period, was not determined. The amounts of NH ⁺₄ -N released in summer vary greatly within the groups of silty (20–200 kg NH ⁺₄ -N ha⁻¹ 30 cm⁻¹) as well as clayey (40–188 kg NH ⁺₄ -N ha⁻¹ 30 cm⁻¹) soils. Studies in which subsoils (>30 cm) were included indicate even higher amounts of NH ⁺_{4 f}-N release. For example, estimates were up to 250 kg NH ⁺_{4 f}-N release to a soil depth of 70 cm in Belgium (Van Praag et al. 1980) and up to 350 kg NH ⁺_{4 f}-N release to a soil depth of 75 cm in Germany (Mengel and Scherer 1981).

Although the total amounts of NH ⁺_{4 f} (“native” plus “recently fixed”) are probably much higher in the clayey soils (clay contents >30%) compared to the silty soils (10–20% clay), the amounts of NH ⁺_{4 f} which are available in a growth period of annual crops seem to be in a similar order of magnitude (Table 3). This indicates a wider ratio of “native” to “recently fixed” NH ⁺₄ in the clayey soils. The low values of NH ⁺_{4 f} release presented in Table 3 can partly be attributed to the fact that the experimental plot was kept fallow (study by Mba-Chibogu et al. 1975) or that the experimental soil had a history of only moderate N fertilization (not more than 60 kg N ha⁻¹ year⁻¹ for 25 years prior to the initiation of the study) with a low yield level (study by Soon 1998). In contrast, soils with a high level of long-term N fertilizer application and a high yield level show a high temporal variation in the interlayer NH ⁺₄ concentration, especially on control plots with limited contents of mineral N (see plot with nil N application in the study by Nieder et al. 1996). Most field observations show that the amounts of NH ⁺₄ -N refixed in autumn are commonly smaller compared to the amounts of NH ⁺₄ -N which are fixed in spring minus the amounts of NH ⁺_{4 f} released in summer. It is thus obvious that for replenishment of the plant-available NH ⁺_{4 f} pool to levels that occur at the beginning of the growth periods, the preceding winter periods would additionally be required.

Fixation is usually faster than release of NH ⁺₄ (Drury and Beauchamp 1991). Kowalenko and Cameron (1976) observed that more than one half of the added NH ⁺₄ -N was fixed within 1.7 days, whereas the average daily release of NH ⁺_{4 f} was 1.7 and 0.65 kg N ha⁻¹ between 19 June to 1 August 1974 and between 1 August to 13 September 1974, respectively (Kowalenko 1978). From the foregoing, it appears that while the native NH ⁺_{4 f} has no significance in the soil N dynamics (Mengel and Scherer 1981; Smith et al. 1994), the temporal changes in the content of recently fixed NH ⁺₄ show that this fraction is actively involved in the N dynamics during the crop growth period. Added NH ⁺₄ is quickly fixed by the clay minerals and later released slowly during the crop growth season due to increased crop demand with concomitant decrease in NH ⁺₄ concentration in soil solution. Supply of C-containing root exudates by the plant enhances the activity of heterotrophic microorganisms which may promote the release of fixed NH ⁺₄ .

The phenomenon of temporary fixation and release of added fertilizer NH ⁺₄ may contribute to retarding nitrification and thus to reducing N losses from the soil–plant system via NO ⁻₃ leaching and denitrification (N₂, N₂O). In the winter-humid temperate climate, it is generally observed that contents of NH ⁺_{4 f} reach their maximum during the winter period, when nitrate leaching occurs frequently. The extent of nitrate leaching may then partly depend on the NH ⁺₄ fixation capacity of the soil. For example, the losses of added fertilizer N (¹⁵NH ⁺₄ ) were more than double in a Histosol with extremely low NH ⁺₄ -fixing capacity as compared to that from a clay containing Gleysol with high fixing capacity (Fischer et al. 1981). Similarly, NH ⁺₄ -N fixation by clay minerals may also contribute to reducing NH₃ volatilization losses (Dou and Steffens 1995). It may, therefore, be argued that clay fixation of NH ⁺₄ can provide a temporary sink for fertilizer N that subsequently acts as a source of N for plant uptake. Information on soil NH ⁺₄ -N fixation capacity and its plant availability is necessary for developing efficient N fertilizer management programs.

Although the knowledge on the magnitude and the dynamics of NH ⁺_{4 f} has increased during the last few decades, there are still conflicting reports about the importance of NH ⁺_{4 f} for plants and microflora. This may be due to the fact that up to now, it is not possible to distinguish properly between native fixed (non-available) and recently fixed (plant-available) NH ⁺₄ . As a consequence, the pool of plant-available NH ⁺_{4 f} has hardly been integrated into models to describe the N dynamics of soils. Further research on methodological aspects is, therefore, required for both proper separation of recently fixed NH ⁺₄ from native fixed NH ⁺₄ and modeling the kinetics of plant-available NH ⁺_{4 f}. This would be a basis for integrating the pool of recently fixed NH ⁺₄ in fertilizer management programs.