ABSTRACT

The dissociation of ammonium into ammonia (NH₃) in wastewaters is a key factor governing atmospheric nitrogen volatilization. Relatively rich in total ammoniacal nitrogen (TAN or plus NH₃), livestock manures are most susceptible to NH₃ volatilization, although indirect measurements report 5 times less dissociation as compared to theoretical values. The objective of this study was therefore to directly measure dissociation of two standard NH₄Cl solutions (1750 and 3500 mg TAN/L), and of swine and dairy cattle manures at various dilutions and temperatures using a ammonia selective electrode (hydrogen and silver-silver chloride electrode couple) at various pH and temperatures. All solutions demonstrated dissociation varying from theory, especially because of dissolved compounds such as atmospheric CO₂. At a neutral pH, ratios of theoretical to measured [NH₃-N] ranged from 1.5 to 3.5, with higher ratios corresponding to higher TAN levels. At a pH below 6, NH₃ volatilization was enhanced by the shift of to H₂CO₃ and CO₂. With previous research projects reporting 5 times less NH₃ volatilization as compared to theory, the present indicates that dissociation activity account for half of this drop with gas diffusion accounting for the other half.

1. Introduction

Worldwide, food production generates 70% of the global ammonia (NH₃) emissions, representing 55 million tons/ yr [1] and being concentrated especially in Europe, the Indian subcontinent, China and the US [2-4]. Within this 70%, fertilizers and livestock manures contribute 10 - 12 and 16 - 27 million tons/yr of NH₃-N volatilization, respectively, where the share of livestock shelters including manure storage and manure land spreading is 30 to 55% and 15% to 40%, respectively [5].

Nitrogen deposition as a result of NH₃-N atmospheric loading leads to several well acknowledged environmental issues: increased risks of water eutrophication [6,7] and acidification [8]; N accumulations changing plant species for soils [9,10], and benthic and pelagic organisms such as corals and algae for waters [11]; increased organic soil subsidence [12]; increased crop sensitivity to secondary stresses such as pathogens, frost and drought [13]; increased CO₂ emissions from forest soils [14], and; damages to tomatoes, cucumbers, conifers and fruit crops [15].

Models simulating NH₃ volatilization from wastewaters rich in such as livestock manure, are based on a dissociation constant pK_a. Nevertheless, available pK_a coefficients pertain to pure solutions rather than natural solutions containing gases dissolved from the atmosphere and minerals in trace quantities. Hashimoto and Ludington [16] estimated the pK_a coefficient of poultry manure slurries by measuring NH₃ volatilization in reaction vessels equipped with a mixing system and assuming a value for Henry’s constant equivalent to that of water. They recommended a manure pK_a coefficient 5 times lower than that computed theoretically for pure water solutions. Similarly for dairy manure, Chaoui et al. [17] observed a pK_a coefficients 80% smaller than those predicted theoretically by measuring the rate of NH₃ volatilization from non stirred jars and once more, assuming a value for Henry’s constant equivalent to that of water. Such low dissociation levels were used by Zhang et al. [18] to produce a model predicting NH₃ losses from manures and by Masse et al. [19] estimating the efficiency of a membrane separation technology.

The assumption that Henry’s constant in manures is the same as that of water may introduce an error in estimating pK_a since the solids content of manures increases their viscous properties and accordingly, decreases their rate of gas diffusion [20]. For anaerobic digesters, Hafner and Bisogni [21] demonstrated that the presence in particular of inorganic carbon could modify pK_a by 20%. To obtain pK_a values without estimating Henry’s constant, NH₃ in solutions must de directly measured. Since the silver-silver chloride electrode, namely the NH₃ selective electrode, is recognized as a reliable instrument to measure NH₃ in solutions [22,23] being sold by several electrode manufacturers, this method is proposed to directly measure NH₃ concentrations and compute measured pK_a coefficients [22,23].

The objective of the present paper was to directly measure the concentration of NH₃ resulting from dissociation at various temperatures and pH, for two NH₄Cl solutions and for fresh swine and dairy cattle manures at different dilution levels. Concentrations of NH₃ ([NH₃-N]) were directly measured using a NH₃ selective electrode connected to a pH meter (Orion, Boston, USA) and converted into a dissociation coefficient pK_a. The two NH₄Cl solutions had a Total Available Nitrogen (TAN or -N plus NH₃-N) of 1750 and 3500 mg/L, and the swine and dairy cattle manures were diluted to various levels with distilled water to offer a TAN in the range of 400 to 3800 mg/L.

2. Chemistry of Ammonium Dissociation

Because of the polar nature of water molecules, H⁺ ions are constantly being exchanged and a self ionization coefficient, pK_w, of more or less 14 is exhibited:

(1)

Pure water exhibits equal concentrations of H⁺ and OH^–, [H⁺] and [OH^–], or a pH (–Log₁₀ [H⁺]) of 7. Ionic elements or compounds dissolved in water take part in the exchange of H⁺, shifting the pH under or above 7, whether an acid or a base generating H⁺ or OH^–, respectively.

The ion is a weak acid dissociating as follows in water:

(2)

The level of dissociation into NH₃ depends on the availability of H⁺ in solution, where a deficit (alkaline conditions) shifts Equation (2) to the right, thus releasing H⁺ from and producing NH₃. The extent of this dissociation process is expressed by the dissociation coefficient K_a, defined as follows:

(3a)

Since K_a coefficients are generally very small, such as that of water at 10^–7, the general practice is to express dissociation coefficients as pK_a where:

(3b)

This principle of dissociation applies to all acids (pK_a) and bases (pK_b) dissolved in water. The higher the value of K_a, or the lower the value of pK_a, the more active is the acid, and the greater is the extent of dissociation exchanging H⁺ ions.

For pure solutions in water, the dissociation coefficient, pK_a, was initially determine in the 1940’s by measuring the electromotive forces created in a cell between a hydrogen and silver-silver chloride electrode couple (Bates and Pinching 1949). Such cell better known as a NH₃ sensitive electrode can produce an error of +/– 5% [24] as a result of variations in the solution solubility of silver chloride and losses of NH₃ during the reading process [23]. The ratio of NH₃ to TAN, f, can be defined in terms of pH [25] based on:

(4a)

The solution to this equation gives (Loehr, 1984):

(4b)

To compute pK_a, the temperature effect can be included as [26]:

(5)

where T is in °K.

Water solutions are rarely pure and pK_a reported for pure TAN solutions cannot be applied. In natural solutions, the extent of the dissociation process depends not only on the solution pH but also the presence of other acids or bases readily exchanging H⁺ without a noticeable change in pH [27]. An important acid affecting [H⁺] in water solution is carbonic acid (H₂CO₃) generated from CO₂ dissolved from the atmosphere or produced from microbial activity degrading organic matter. Carbonic acid releases H⁺ as the pH increases or [H⁺] decreases, with the maximum and occurring at a pH of 6.5 and 10.5, respectively [28,29]:

(6)

By releasing H⁺, creates a buffering effect and increases the amount of available H⁺, thus potentially decreasing the dissociation of into NH₃. Furthermore is the preferred ionic form as compared to, with a lower pKa coefficient (Table 1).

Besides dissolved atmospheric and microbial CO₂, manures offer a wide range of compounds interfering with H⁺ exchanges, such as phosphates, with maximum concentrations of and occurring at a pH of 5 and 9.5, and with pK_a of 2.13, 7.20 and 12.37 for, and respectively (Table 1):

(7)

As compared to, both H₃PO₄ and H₂CO₃ have smaller pK_a coefficients, with a greater potential of governing the availability and exchange of H⁺.

In conclusion, the dissociation coefficients expressed by pKa developed for pure solutions of specific acids, including, cannot be directly applied to natural solutions because of the presence and interaction of other more active species. Compared to water solutions with a limited amount of dissolved ions and solids, wastewaters such as manure can present a more complex reaction affecting dissociation, thus the need to conduct direct measurements.

2. Materials and Methods

2.1. Experimental Solutions

Fresh dairy cattle and swine manures were collected from the Dairy Cattle and Swine Complexes of the Macdonald Campus Research Farm of McGill University, located at the western tip of the island of Montreal, Canada. The samples were transported to the laboratory and refrigerated at 5˚C to 6˚C until used. Each sample was characterized in triplicate for total solids (TS), total volatile solids (TVS), pH, total Kjeldhal nitrogen (TKN),

TAN, total phosphorous (TP) and total potassium (TK), according to standard methods (APHA, 2005). To conduct the NH₃ trials, swine manure was tested whole and after being diluted with distilled water, while dairy cattle manure, harder to stir because of its fibrous solids, was only tested at a dilution ratio of 1:4. The NH₄Cl salt used to prepare the standard solutions was purchased from Fisher Scientific (98% purity).

2.2. Methodology

The NH₃ sensitive electrode is a well recognized method for the measurement of TAN concentrations in solutions and manures. Generally, the pH is adjusted to 13, to dissociate practically all into NH₃. Therefore, this project use an NH₃ sensitive electrode as follows to directly measure [NH₃-N] [24]: 1) the pH of a standard NH₄Cl solution or manure sub-sample was increased to 13 using 5 M NaOH and its TAN (total NH₃ and) was measured; 2) another sub-sample of the same NH₄Cl solution or manure sample was tested for NH₃-N after gradually changing its pH no more than 4 times, with either HCL or NaOH, while also measuring its temperature; 3) the process was repeated at various pH and temperatures. The measured NH₃-N/TAN fraction f was used to compute the measured pK_a coefficient for a given pH and temperature as follows:

(8)

The temperature of the solutions and manures was modified by storing in a closed flask in a refrigerator or leaving on a bench top for 24 h to obtain room temperature. The temperature of the solution or manure was measured using a standard mercury laboratory thermometer with an accuracy of +/– 0.5˚C. Solutions of 1 M HCl and 5 M NaOH were used respectively to drop and increase the pH.

The procedure described above was initially used to quantify [NH₃] in standard NH₄Cl solutions of 3500 and 1750 mg TAN/L, at a pH increased from 6 to 13 and a temperature ranging from 15˚C to 25˚C. For each pH and temperature combination, the procedure was repeated 3 times.

The manure samples were tested at various pH using the same procedure. Swine manure with a 8.8% total solids (TS) was tested at a temperature and dilution ratio with distilled water of: 20˚C - 25˚C at a dilution of 0 (1 part manure and 0 part distilled water); 8˚C - 15˚C and 20˚C - 25˚C at a dilution ratio of 1 (1 part manure and 1 part distilled water), and; 8˚C - 15˚C and 20˚C - 25˚C at a dilution ratio of 4 (1 part manure and 4 part distilled water). The procedure was repeated with 9.9% TS cattle manure at temperatures of 8˚C to 25˚C, diluted with distilled water at a ratio of 4 (1 part manure and 4 part distilled water). All measured NH₃ concentrations, pK_a, coefficients and f fraction were compared to those obtained theoretically using Equations 4b.

2.3. Analytical Procedure

Both the swine and dairy manure samples were analyzed using standard methods (APHA 2005) for total solids (TS), and total volatilize solids (TVS), pH, total Kjeldhal nitrogen (TKN), total phosphorous (TP) and total potassium (TK). Total solids were determined by drying at 103˚C for 24 h in an oven (Scientific John by Sheldon Manufacturing Inc., Cornelius, Oregon, USA). The TVS were determined by burning the dried samples for 4 h in a muffle furnace (Blue M Electric Company, Blue Island, Illinois, USA). The pH was determined using a pH probe attached to a pH/Ion meter (Orion Model 450, Boston, Massachusetts, USA).

The TAN value was measured by increasing the pH of the solution to 13 using 5 M NaOH, and using an NH₃ selective electrode (Orion 9512HBNWP ammonia sensitive electrode, Boston, USA) for quantification with an accuracy of 2% of the reading. Total Kjeldhal nitrogen (TKN) was determined after digesting the samples in 18 M sulfuric acid and 50% hydrogen peroxide for 15 minutes at 500˚C. The TKN of the digested samples was determined at a pH of 13, using an NH₃ selective electrode connected to a pH meter. The electrode was calibrated before hand using standard NH₄Cl solutions. The TP and TK of the digested samples were quantified at a pH of 7, using a spectrophotometer (Hach, Model DR 5000, Loveland, Co., USA).

2.4. Statistical Analysis

Standard deviations were obtained using Excel (Microsoft Office 2007, Seattle, USA). The ANOVA was used to measure differences between treatments (solution type and temperature), where pH range was used as block.

3. Results and Discussion

3.1. Characteristics of the Experimental Manures

The characteristics of the experimental cattle and swine manures are described in Table 2. The two manures had similar TS with that of dairy cattle and swine offering 9.9 and 8.8%, respectively. The TKN and TAN concentrations differed between the dairy and swine manure, with the dairy manure offering concentrations of 3720 and 2170 mg/L, and the swine manure offering concentrations of 5500 and 3825 mg/L, respectively. The pH of both manures was in the neutral range of 7.4 and 7.1 respectively. The total phosphorous and potassium content of the swine manure were higher than those of the dairy

cattle manure at 2020 and 3700 mg/L, compared to 700 and 2600 mg/L.

3.2. Ammonium Dissociation for NH₄Cl Water Solutions

The pH of the NH₄Cl solutions at various temperatures ranging from 15˚C to 25˚C is the first interesting observation. At 15˚C and 25˚C and in pure water, exhibits a pK_a of 9.55 and 9.24 computed from Equation 5. Accordingly, 0.25 and 0.125 molar solutions (3500 and 1750 mg TAN/L) should produce a pH of 4.9 and 5.1, respectively. For the solution at 0.25 molar (3500 mg/L TAN), the observed pH ranged from 5.6 to 5.9 irrespective of the temperature ranging from 15˚C to 20˚C. Another series of tests produced a pH of 5.9 to 6.0 at 25˚C for the 0.125 and 0.25 molar (1750 and 3500 mg/L TAN) solutions.

The discrimination between the measured and expected pH therefore resulted from the interference of other dissolved compounds or elements at the time of testing, besides the effect of NH₄Cl. Because the observed pH was higher than that expected theoretically, the dissolved compounds or elements were either absorbing the H⁺ ions generated by the NH₄⁺ dissociating into NH₃, or simply attenuating the extent of the dissociation process as a result of a higher H⁺ availability. Since the NH₄Cl solution was prepared in distilled water, dissolved atmospheric CO₂ with a pK_a of 6.35 is the most logical compound capable of producing such interference. Since the measurements were conducted in January and February, the ambient laboratory air likely contained CO₂ levels of 800 to 1000 ppm, standard winter CO₂ concentrations in public buildings. These concentrations enhanced the amount of dissolved CO₂ and accordingly H₂CO₃. Accordingly, the initial solution pH indicated that dissociation was influenced by other dissolved compounds primarily dissolved from the atmospheric, such as CO₂ forming absorbing H⁺ and buffering a change in pH with the addition of NH₄Cl.

The ratio of theoretical to measured [NH₃] is presented in Figure 1(a), for both NH₄Cl solutions, as a function of pH. The first observation is that, below a pH of 6.5, the 3500 mg TAN/L at 15˚C - 20˚C and 1750 mg TAN/L at 25˚C produced a ratio generally under 1.0, except for the 3500 mg TAN/L solution at 25˚C which produced a ratio between 0.9 and 1.5. At a pH above 6.5, both the 3500 mg TAN/L produced a peak ratio of 2.3, as compared to the 1750 mg TAN/L producing a peak ratio of 1.5. At 25˚C, and compared to the 3500 mg/L solution, the 1750 mg/L solutions produced a lower theoretical to measured [NH₃] ratio for a pH range of 7.7 to 8.7 (95% probability).

The observed ratio of theoretical to measured [NH₃] in Figure 1(a) is explained by the buffering action of H₂CO₃. When adding a weak acid such as to a solution of, the reaction illustrated by Equation 6 shifts to the left, thus producing H₂CO₃ absorbing H⁺, creating a H⁺ deficit and increasing [NH₃]. Furthermore, this left shift of reaction for Equation 6 releases CO₂ to the atmosphere further increasing the deficit of H⁺. This

phenomenon is also observed in high salt solutions where CO₂ solubility is reduced. Finally, this left shift of reaction for Equation 6 results in the absorption of H⁺ ions, explaining the higher observed pH discussed above. For the 3500 mg TAN/L solution at 25˚C with a higher ratio of theoretical to measured [NH₃] as compared to the other treatments, the explanation is a higher concentration versus a lower dissolved CO₂ presence, thus shifting the H⁺ balance. At 25˚C, CO₂ is less soluble.

At a pH above 6, the [OH^–] increases and the reverse reaction happens. The measured [NH₃] is lower than that expected theoretically because adding OHto an existing solution of and shifts the reaction of Equation 6 to the right. Thus, dissociation is reduced as compared to theoretical computations because of a lower deficit of exchangeable H⁺ [29]. Interestingly enough, temperature resulting in a lower CO₂ concentration and concentration had less effect at such a pH range. Nevertheless, the 1750 mg TAN/L solution produced a significantly higher (95% confidence level) percentage of dissociation for a pH range of 7.5 to 9.0, likely because of its lower concentration requiring less H⁺. Instead of producing twice the NH₃, the 3500 mg TAN/L solution produced only 10% more (6 × 10^–9 moles of NH₃ as compared to 5 × 10^–9 moles of NH₃) as compared to the 1750 mg/L solution at a pH of 7.5 to 9.0.

The ratio of theoretical to measured pK_a as a function of pH range is illustrated in Figure 1(b) with error bars based on the coefficient of variance. As compared to Figure 1(a), this comparison better illustrates temperature effects for the 3500 mg TAN/L solution below a pH of 6.5, with more dissociation at 25˚C. At a pH above 6.5, the 1750 mg/L solution produced a significantly lower ratio of theoretical to measure pK_a as compared to the 3500 mg TAN/L solution (95% probability). For the 3500 mg TAN/L solution at 15 - 20 and 25˚C, and the 1750 mg TAN/L solution at 25˚C, the measured pK_a was reduced by 9.65, 9.69 and 9.81 respectively, as oppose to that computed theoretically.

In conclusion and at neutral pH range, simple NH₄Cl solutions do not dissociate according to the theoretical pK_a coefficient associated with pure solutions. Rather and because of the presence of especially dissolved atmospheric CO₂, the NH₄Cl solutions demonstrated a lower dissociation level compared to theoretical values, by a factor of 1.5 to 2.3, for TAN concentrations of 1750 to 3500 mg/L. Nevertheless, at a pH under 6 and a temperature of 25˚C, such solutions tended to demonstrate a higher dissociation levels as compared to the theoretical values for pure solutions.

3.3. Ammonium Dissociation with Swine and Cattle Manures

The swine manure was tested at different temperatures and dilutions: 20˚C to 25˚C for a dilution of 0 (1 part manure to 0 part distilled water); 8 to 15 and 20˚C to 25˚C for a dilution of 1 (1 part manure to 1 part distilled water), and; 8 to 15 and 20˚C to 25˚C for a dilution of 4 (1 part manure to 4 parts distilled water). The four dilutions produced 3800, 1910 and 765 mg TAN/L, where the first two dilutions had a TAN equivalent to that of the two NH₄Cl solutions.

Because manures contain a much wider range of compounds and solids, their dissociation as a function of pH should be more complex and variable than that of the NH₄Cl solutions. Manures contain: solids affecting viscosity and therefore the rate of exchange of CO₂ with the atmosphere; organics generating dissolved gases such as CO₂, H₂S and CH₄ as a result of microbial degradation, and; soluble compounds including organic acids and H₃PO₄ [20,21] with a lower pK_a than that of. For example, acetic acid has a pK_a of 4.75 compare to that of at 9.25.

In Figure 2(a), the theoretical to measured [NH₃] is illustrated as a function of pH. For all three dilutions, the trend is similar to that of the NH₄Cl solutions. Nevertheless, some differences are observed. Below a pH of 6.0 and for the dilutions of 1 and 4, the ratio of theoretical to measured [NH₃] is lower than that of the NH₄Cl solutions, likely because of the additional CO₂ produced by the microbial activity also know to increase with manure dilution. Still as compared to the NH₄Cl solutions and for a pH of 6 to 8, the ratio of theoretical to measured [NH₃] reaches a higher peak of 2.4 to 3.5, with higher values associated to the higher TAN swine manures with lower dilution rate.

At a pH under 6.0, higher CO₂ levels generated from microbial activity and lower TAN concentrations for the diluted manures push the ratio of theoretical to measured [NH₃] towards lower numbers below 1.0. This ratio for the less diluted manure at 3800 mg TAN/L remained generally above 1.0, similarly to the 3500 mg TAN/L solution of NH₄Cl. For a pH of 6 to 9, the swine manure diluted 4 times tended to present a theoretical to measured [NH₃] ratio of more or less 1.0. This can simply result from its dilution producing a lower TAN concentrations requiring less H⁺ for the dissociation process.

Still for the swine manure, the ratio of theoretical to measured pK_a as a function of pH is illustrated in Figure 2(b) with a trend similar to that of the [NH₃] ratio. On the average and for a pH between 6.5 and 9, the measured pK_a for the swine manure diluted to a level of 0, 1 and 4 produced a pK_a equal to 0.96, 0.975 and 0.995 times that computed theoretically. For all pH tested, the ratio of theoretical to measured pK_a was inversely proportional to the swine manure TAN. Because of the result variability caused by the presence of other compounds, the difference was significant between the 0:1 and 4:1 dilutions

(95% confidence level) only for a pH of 5.0 to 6.0.

The ratio of theoretical to measured [NH₃] of the dairy manure diluted 4 times is illustrated in Figure 3(a) at a TAN of 400 mg/L. This ratio follows the trends observed with the swine manure and the NH₄Cl solutions. Nevertheless, the 8˚C to 15˚C tests provided a linear relationship with pH, on the average and despite the result variation, whereas the 25˚C test demonstrated a ratio decreasing from 2.6 to 1.2, for a pH of 7 to 9. This may result from a more active microbial population at 25˚C since the samples were left on the bench top for 24 h before testing. Cattle manure has higher fiber content than that of swine manure, offering a lower and more variable microbial digestibility especially at low temperatures.

For the diluted cattle manure, the ratio of theoretical to measured pK_a as a function of pH is illustrated in Figure 3(b). This figure demonstrates the temperature effect on the pK_a ratio, especially for a pH of 7 to 8. On the average and for a pH between 6.5 and 9, the measured pK_a for the dairy cattle manure was equal to 0.975 and 0.98

times that computed theoretically.

4. Conclusions

The objective of the present research was to measure the level of dissociation at temperature of 10˚C to 25˚C, for two NH₄Cl solutions, and for swine and dairy cattle manure at different dilutions with distilled water. Ammonia concentrations ([NH₃]) were measured using an NH₃ selective electrode consisting of a hydrogen and silver-silver chloride electrode couple connected to a pH meter. The theoretical and measured [NH₃] and dissociation coefficient, pK_a, were compared as a function of pH and temperature, respectively.

The results obtained with the two NH₄Cl solutions indicated that dissociation is highly influenced by dissolved compounds, especially atmospheric CO₂ with a lower pK_a coefficient and therefore a greater affinity for H⁺. Dissolved CO₂ had the effect of increasing dissociation at a pH under 6, but decreasing its dissociation at a pH of 6 to 10, resulting in a measured pK_a higher than that computed theoretically. Lowering the TAN concentration of the solution produced a measured [NH₃] approaching that computed theoretically because of a lower H⁺ demand.

For the swine and dairy cattle manures, the ratio of theoretical to measured [NH₃] was also influenced by the TAN concentration or the level of dilution with distilled water. The more diluted manure showed a dissociation activity approaching that computed theoretically. Nevertheless, the undiluted swine manure produced a ratio of theoretical to measured [NH₃] of 3.3 at a neutral pH. For the two solutions and manures tested, temperature within a range of 10˚C to 25˚C, had little if any effect on dissociation for a normal pH range of 6.5 to 9.

The present research therefore indicates that manures and solutions, exposed to atmosphere and at a neutral pH, show less dissociation activity than that expected theoretically, by a factor ranging from 1.5 to 3.3. This factor is influenced by the TAN concentration of the manure or solution, with lower TAN levels producing values closer to those computed theoretically. With previous research projects reporting 5 times less NH₃ volatilization as compared to theoretical computations, the present indicates that this phenomena results from a dissociation activity generally 2.5 times lower, and thus, a gas diffusion process 2.0 times slower.

5. Acknowledgements

This project was funded by the Natural Science and Engineering Research Council of Canada.

REFERENCES

NOTES

^*Corresponding author.