Technical TECHNICAL reports: Waste REPORTS Management
Maximizing Ammonium Nitrogen Removal from Solution Using Different Zeolites Chad J. Penn,* Jason G. Warren, and Savannah Smith Oklahoma State University
Zeolite minerals are ideal for removing ammonium nitrogen (NH4+–N) from animal wastes, leachates, and industrial effluents. The objectives of this study were to compare NH4+ removal and kinetics among several commercially available zeolites under various conditions and determine if calorimetry could provide information regarding kinetics of NH4+ removal. Ammonium sorption onto potassium (K) saturated zeolites was compared using synthetic vs. natural swine effluent and with either traditional batch-shaken system or a “tea bag” approach in which zeolites were contained in a mesh sack and suspended in a solution of swine effluent. Ammonium sorption was measured at four retention times using a flowthrough system, and the resulting heat response was measured using isothermal calorimetry. Ammonium removal was not significantly different in synthetic vs. natural swine effluent. Ammonium removal was lower in batch-stirred compared to batch-shaken systems, suggesting that diffusion between particles was rate-limiting in the former system. Flow-through cells possessing contact times >100 s displayed greater NH4+ sorption than batch systems, suggesting that maintaining high NH4+ concentration in solution, removal of exchange products, and sufficient reaction time are critical to maximizing NH4+ removal by zeolites. Within 100 s after NH4+ addition, endothermic heat responses indicated that NH4+–K+ exchange had peaked; this was followed by significant heat rate reduction for 50 min. This confirmed findings of an initial fast NH4+– K+ exchange followed by a slower one and suggests the 100-s period of rapid reaction is an indicator of the minimum flow through retention time required to optimize NH4+ sorption to zeolites used in this study.
Copyright © 2010 by the American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America. All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. J. Environ. Qual. doi:10.2134/jeq2009.0324 Published online 11 May 2010. Received 21 Aug. 2009. *Corresponding author (
[email protected]). © ASA, CSSA, SSSA 5585 Guilford Rd., Madison, WI 53711 USA
A
gricultural and industrial effluents rich in ammonium nitrogen (NH4+–N) can sometimes negatively impact water quality on discharge to surface waters or soils (Carpenter et al., 1998). Discharge of NH4+ rich effluents is common among swine production operations, dairy processors, oil refineries, distilleries, domestic wastewater treatment plants, and fertilizer plants. For high pH effluents such as swine effluent, even storage of the liquid itself poses problems since a large percentage of NH4+ is lost to the atmosphere as ammonia (NH3) gas (Panetta et al., 2005). Ammonia gas is considered an air pollutant and can also be deposited directly into surface waters during precipitation events (Wu et al., 2003). Losses of N to the environment are not only a potential environmental problem, but discharge of this valuable plant nutrient represents a monetary loss. Zeolites have been shown to remove a high amount of NH4+ from solutions via ion exchange due to their high permanent negative charge density and the presence of inner “channels” for ion diffusion (Ames, 1967; Weber et al., 1983; Rozic et al., 2000; Zaman et al., 2007, 2008). Weber et al. (1983) amended a Nunn clay loam (fine, smectitic, mesic Aridic Argiustolls) with zeolite and significantly reduced the leaching of NH4+ compared to unamended soils. Zeolites have also been used in simulated sandbased golf greens to successfully reduce NH4+ and NO3- leaching (Huang and Petrovic, 1994) and also for treatment of dairy effluent before discharge (Bolan et al., 2004). Other studies have examined the kinetics of NH4+ removal from solutions (Kithome et al., 1998; Cooney et al., 1999; Nguyen and Tanner, 1998; Zhao et al., 2004). However, most of these studies have focused on batch (i.e., static) systems with fewer studies conducted using a flow-through approach. For example, Kithome et al. (1998) used a batch system to examine the kinetics of NH4+ removal from the zeolite, clinoptilolite, and found that first-order kinetics, Freundlich, parabolic diffusion, Elovich, and heterogeneous diffusion models adequately described the sorption process. Although batch kinetics data are insightful, a flow-through technique is more representative of natural processes (i.e., surface runoff of NH4+–N, leachate passing through a soil) and industrial applications (i.e., NH4+ stripping systems). Using a flow-through approach, Cooney et al. (1999) examined the kinetics of zeolite NH4+ removal from a municipal sewage effluent, and showed that zeolites exhibited high NH4+ removal that was sufficient to support a continuous high rate process. However, the authors did Dep. of Plant and Soil Science, Oklahoma State Univ., 367 Agricultural Hall, Stillwater, OK, 47078-1020. Assigned to Associate Editor Nanthi Bolan. Abbreviations: CEC, cation exchange capacity; DI, deionized; EC, electrical conductivity; ITC, isothermal titration calorimetry.
in size to NH4+ and will therefore easily displace sorbed NH4+ when added to zeolites at a high concentration. Swine effluent was collected from the Oklahoma State University swine lagoon and analyzed for Ca, K, Cu, Mg, P, and S using inductively coupled plasma atomic emission spectroscopy (ICP–AES). Nitrate (NO3-) and NH4+ were analyzed colorimetrically (LACHAT, 1994). Effluent pH and electrical conductivity (EC) were determined by pH and EC electrode, respectively. Total N and carbon (C) were determined on a LECO Truspec dry combustion C analyzer (Nelson and Sommers, 1996). A synthetic effluent was made that possessed the same chemical constituency as the natural swine effluent, with the exception of organic compounds. The synthetic effluent was made with ammonium nitrate and chloride and sulfate salts. Solution pH was adjusted with 5 M NaOH.
not test the effect of different flow rates (i.e., contact time) on NH4+ removal. Nguyen and Tanner (1998) conducted a study using two different zeolite minerals with varying particle distribution to examine NH4+ removal kinetics in dairy, swine, and municipal effluents. Although two different contact times were examined, both were relatively long (0.32 and 17.78 h). Thus, there remains a need to examine zeolite NH4+ removal kinetics in systems with shorter contact times (i.e., faster flow rates) since long retention times may not be economically feasible or practical in all cases. One tool that may aid in such an objective is isothermal titration calorimetry (ITC). This technique allows one to precisely monitor changes in heat on titration of a known solution to a solid material. The degree and timing of the change in heat (either exothermic or endothermic) can serve as an indicator of the extent to which the reaction has taken place. For example, Kallay et al. (1997) used calorimetry to determine the rate constant and enthalpy of dissolution of solid calcium carbonate in aqueous ethylenediaminetetraacetic acid (EDTA) and NaOH solutions. Similarly, Rhue et al. (2002) conducted a calorimetry study on potassium (K+) and calcium (Ca2+) exchange on a dowex exchange resin and an Oxisol. These authors found that the thermograms provided valuable information regarding the kinetics of ion exchange; specifically, Ca2+ displaced K+ faster than the reverse exchange process for the resin. The objectives of this study were (i) to compare NH4+ removal and the impact of the solid:solution ratio for several commercially available zeolites, (ii) to determine kinetics and efficiency of NH4+ removal in swine effluent using a “tea bag” approach in which a permeable sack containing zeolites was suspended in the batch system, (iii) to determine the effect of flow rate (contact time) on NH4+ removal in a flow-through system, and (iv) to determine if ITC could potentially serve as an indicator of reaction kinetics for NH4+ sorption.
Synthetic effluent was reacted with zeolites A, B, C, and D (replicated three times) at a 1:20 solid:solution ratio for 1 h, followed by centrifugation for 10 min at 1200 rpm. The supernatant was then separated and analyzed for NH4+ colorimetrically. The amount of NH4+ sorbed was calculated from the difference between the NH4+ concentration in the synthetic effluent and equilibrated solution. The same zeolite samples were then exposed to a 1 M KCl solution (1:20 solids:solution ratio) for 1 h to estimate exchangeability of the sorbed NH4+. Samples were centrifuged and analyzed for NH4+ as previously described. On the basis of these results, zeolites A and D were then utilized for further experimentation. Batch-shaken experiments were conducted on zeolites A and D using natural swine effluent. Samples were equilibrated in the exact same manner as described for the synthetic effluent. In addition, several solid:solution ratios were tested (n = 3: 1:40, 1:20, 1:10, 1:5, 1:2, and 1:1). Samples were equilibrated and analyzed the same way as previously described.
Materials and Methods
Batch-Stirred Kinetics
Four commercially available zeolites were obtained from Bear River Zeolite Corporation, (Zeolite A; Thompson Falls, MT), Steel Head Specialty Minerals (Zeolite B; Spokane, WA), and Zeotech Corporation (Zeolites C and D; Fort Worth, TX). More specifically, zeolites C and D are known as “Zeomax SC” and “Zeomax LM,” respectively. The zeolite mineral contained in all products was clinoptilolite; the percentage purity and total surface area are as reported by the respective company. Bulk density was determined on each sample by weighing a known volume of dried material (65°C for 48 h). Cation exchange capacity (CEC) was estimated by the unbuffered salt extraction method (Sumner and Miller, 1996). All materials were saturated with 1 M KCl before all experiments (1 h equilibration time at 1:10 solid:solution ratio) and then rinsed with deionized (DI) water to “normalize” the zeolite surfaces by saturating them with K+ only since different cations have varying degrees of preference for exchange with NH4+ on zeolites (Weatherley and Miladinovic, 2004). Potassium was chosen due to its availability in large quantities, monovalent charge, use as a plant macro nutrient, and nonpollutant in regard to water quality. Potassium is also very similar
Zeolites A and D (25 g) were placed into woven nylon mesh (307 × 307 mesh, 0.005 cm opening) sacks and suspended into 1 L of swine effluent (1:40 ratio). Containers were covered and placed onto a magnetic stirring plate. Five-milliliter effluent samples were removed at time intervals of 0, 1, 3, 5, 10, 15, 30, 60, 120, 240, 360, and 1020 min. A blank consisting of effluent without zeolites was also included. All treatments were replicated three times. Samples were analyzed for NH4+ as previously described. Blank NH4+ values were subtracted from the zeolite treatments when determining sorption values.
Batch-Shaken Experiments
Flow-Through Kinetics Zeolites A and D (5 g) were placed in a flow-through cell (4 cm diameter) possessing a mesh screen and 0.45-mM filter membrane at the outlet. A constant head Mariotte bottle was used to supply synthetic effluent at a depth equal to the surface of the zeolites. Constant flow rates were maintained by attaching the outlet of each flow-through cell to a 12 channel peristaltic pump and adjusting to the desired flow rate. Four flow rates were tested: 4.5, 2.5, 1.0, and 0.5 mL s-1. Aliquots of eluted sample were collected every 15 min for the first 30 min and Journal of Environmental Quality • Volume 39 • July–August 2010
then every 30 min for the remaining 5 h. All solutions were analyzed for NH4+ as previously described.
appreciably in total surface area and percentage purity (i.e., percentage clinoptilolite; Table 2), likely due to variability among different zeolite deposits. Note that zeolites A and D possessed the highest percentage of clinoptilolite and greatest CEC. Zeolites A and D displayed CEC similar to that measured by Cooney et al., 1999 (128 cmolc kg-1), Weber et al., 1983 (140 cmolc kg-1) and Kithome et al., 1998 (93 cmolc kg-1). The relatively low CEC measured for zeolites B and C may be due the lower percentage purity of clinoptilolite contained in those samples (Table 2).
Isothermal Titration Calorimetry All ITC experiments were conducted on a CSC 4200 Isothermal Titration Calorimeter (CSC Inc., Lindon, UT) at 25°C. The ITC had a sensitivity of 0.1 mcal detectable heat effect and a “noise level” of ±0.01 mcal s-1 (deconvoluted signal). All titrations included a single titration of 250 mL of 0.7 M NH4Cl into 0.1 g of zeolite A and D (four replications) suspended in 750 mL of DI water. After titration, Dheat was monitored for 5 h. A blank was determined by titrating the 0.7 M NH4Cl into DI water under the same conditions employed for the actual experiment. Data produced from the blank titrations were subtracted from the sample titrations. Reaction vessels were stirred during all ITC experiments.
Batch-Shaken Experiments When exposed to the synthetic swine effluent solution, all four zeolites sorbed an appreciable amount of NH4+ relative to that reported for soils (Matschonat and Matzner, 1995; Fernando et al., 2005). Zeolites A, B, C, and D sorbed 6.5, 3.8, 3.8, and 7.2 g NH4+ kg-1, respectively (Table 2), which is typical for zeolites tested in batch systems (Nguyen and Tanner, 1998; Weatherley and Miladinovic, 2004; Rozic et al., 2000; Bernal et al., 1993; Weber et al., 1983). Zeolites A and D removed more NH4+ from solution compared to zeolites B and C because they had greater CEC, likely due to possessing a higher percentage purity of clinoptilolite (Table 2). Note that only 32 to 34% of the CEC was occupied with NH4+, suggesting that chemical conditions such as NH4+ concentration, solid:solution ratio, pH, and presence of competing ions will have an impact on NH4+ removal. Ammonium sorbed on zeolites in the synthetic effluent was mostly exchangeable since extraction of the samples with 1 M KCl removed (i.e., exchanged) 81 to 87% of the previously sorbed NH4+. Kithome et al. (1998) showed that 2 M KCl was sufficient for removing 65 to 90% of NH4+ sorbed to zeolite minerals. This high rate of desorption of NH4+ sorbed to zeolite is important for several reasons. In the context of plant growth, the sorbed NH4+ must be plant available when soils are amended with zeolites. Regarding industrial processes or NH4+ removal for environmental purposes, the high exchangeability– desorbability of NH4+ ensures that the zeolite can be reused, which represents a large cost savings.
Statistical Analysis All statistical procedures were conducted using the SAS software (SAS Institute, 2003). Ammonium sorption by zeolites was compared using analysis of variance (ANOVA) at a P value of 0.05 using the PROC GLM command. The PROC NLIN command was used to further examine the relationship between time and cumulative NH4+ removed in regard to flowthrough experiments. The analysis estimated the significant “breakpoint,” or x axis variable (i.e., time) in which the slope significantly changed for the relationship. This also estimated the slope of the line before and after the breakpoint.
Results and Discussion Background Swine Effluent and Zeolite Properties The majority of the N in the swine effluent utilized in the experiments was in the form of NH4+ (Table 1). The difference between total N and NH4+ + NO3- was assumed to be organic N. This measured concentration of NH4+ is considered typical for swine effluent (Hamilton et al., 1997). The four commercially available zeolites all contained the mineral clinoptilolite and were of similar bulk density with the exception of zeolite D (Table 2). However, the zeolites varied
Table 1. Composition of natural swine effluent used in the batch sorption and kinetics experiments. NH4+–N
NO3—N
Ca
K
Cu
Na
Mg
P
S
Total N
————————————————————————— mg L ————————————————————————— 785 0.71 117 529 0.03 167 18 30 27 900 –1
Total C
pH
EC†
—— g L —— 2.3 8.1
m 1646
–1
† EC, electrical conductivity. Table 2. Key properties of zeolites used in the study. Property % Clinoptilolite Total surface area (m2 g–1) Bulk density (g cm-3) CEC (cmolc kg–1)† NH4+ sorbed (g kg–1)‡ % NH4+ desorbed§
Zeolite A 85 25 0.96 108 6.5 81
B 73 40 0.97 65 3.8 84
C 70 13 0.90 66 3.8 83
† CEC = cation exchange capacity; LSD = 3.3 (P < 0.05). ‡ NH4 sorption at a 1:20 solid:solution ratio in synthetic swine effluent, 1-h equilibration time. LSD = 0.45 (P < 0.05). § Percentage of the total NH4 that was previously sorbed. LSD = 11.9 (P < 0.05).
D 80 32 0.84 127 7.2 87
Repeating the sorption experiment using natural swine effluent at the same solid:solution ratio resulted in similar NH4+ sorption compared with the synthetic effluent (averages of 6.3 and 7.6 g kg-1 for zeolite A and D, respectively; Fig. 1, 1:20 solid:solution ratio). This was not expected since natural swine effluent contains many dissolved organic compounds of humic nature (Seol and Lee, 2000) that could possibly coat zeolite particles or compete with NH4+. A hypothesis to explain similar NH4+ sorption among synthetic and natural swine effluent is that the dissolved organic compounds in the swine effluent were large relative to the internal zeolite channels. The largest of such channels are often 0.44 by 0.72 nm for clinoptilolite (Boettinger and Fig. 1. Sorption of NH + in swine effluent onto zeolite under different solid:solution ratios. Vertical 4 Ming, 2002). For example, Ming and bars represent standard deviation (n = 3). Left y axis presents NH4+ sorption on a unit mass basis, and right y axis provides the percentage NH4+ removed in the solution. Dixon (1987) showed that tert-butyl+ ammonium [(CH3)3CNH3 ] was highly Figure 2 shows that sorption was time dependent for this restricted from entering the interior particular approach and far from complete at 1 h. In fact, only + compared with NH4 . 50% of the NH4+ that sorbed in 17 h was completed after 1 As expected, increasing the solid:solution ratio resulted in h. For zeolites A and D, 20 and 33% of total NH4+ sorption the removal of a greater percentage of the swine effluent NH4+ occurred during the first 5 min, indicating that there was an from solution, but the sorbed NH4+ was not as concentrated initial fast reaction taking place. + (i.e., less NH4 sorbed per unit mass zeolite) on the zeolite (Fig. Using NH4+ concentrations appreciably lower than the + 1). A 1:1 ratio removed nearly 100% of the NH4 from solupresent experiment, Kithome et al. (1998) found that the tion. Depending on the scenario, a small solids:solution ratio NH4+ sorption process onto the zeolite clinoptilolite was + may suffice for treating NH4 rich effluents; for example, if the completed at 60 min for a batch-shaken system. However, intention is to produce a more highly NH4+ saturated zeolite when the authors tested a higher concentration of NH4+ for later use as a crop N source. On the other hand, a 1:1 ratio solution (840 mg NH4+ L-1), the sorption process did not would be required if the ultimate goal was to remove the maxiappear to reach equilibrium until about 80 min. Similar to mum amount of NH4+ from solution. Similar to the previous our results, Kithome et al. (1998) found that NH4+ sorption + results, zeolite CEC was not occupied with NH4 only, even at pH 7 was 50% completed at 5 min, relative to the total at the 1:40 solid:solution ratio. More specifically, 67 and 57% sorption at 120 min for a batch-shaken experiment. Other of the measured CEC for zeolites A and D, respectively were studies have shown that NH4+ sorption was completed in 60 + filled with NH4 .
Batch-Stirred Kinetics Ammonium sorption from natural swine effluent during the batch-stirred experiment was less than the batch-shaken test at the same solid:solution ratio (1:40) (Fig. 1 and 2). After 1 h of exposure to the effluent, zeolites A and D only sorbed 4.4 and 5.7 mg NH4+ kg-1, respectively. The observation that NH4+ sorption was less for the batch-stirred than for batch-shaken suggests that exposing zeolites to effluent in the “tea bag” approach physically limits diffusion and flow of the effluent and reaction products (K+) through the pore space between particles compared with shaking a batch of zeolites. Although the suspended “tea bag” approach may be more feasible, simple, and inexpensive than mixing zeolites in a solution for certain scenarios, the data demonstrates that it is a less effective NH4+ removal system. Even after 17 h, the batch-stirred system was not able to sorb as much NH4+ as the batch-shaken system at the same solid:solution ratio in 1 h (Fig. 1 and 2).
Fig. 2. Sorption of NH4+ onto zeolite during17 h using a “tea bag” of zeolite suspended in a stirred vessel of swine effluent (1:40 solid:solution ratio). Journal of Environmental Quality • Volume 39 • July–August 2010
to 120 min (Bernal et al., 1993; Rozic et al., 2000; Weber et al., 1983).
Flow-Through Comparison of flow-through sorption with the batch-sorption results is valid at equal NH4+ loading rates and solid:solution ratios. For example, the addition of 40 g NH4+ kg-1 zeolite by flow-through (1.0 and 2.5 mL min-1) resulted in sorption of 10 to 19 g NH4+ kg-1 (Fig. 3) compared with the batch-shaken test results of 12.5 and 12.9 g NH4+ kg-1 for zeolites A and D, respectively (Fig. 1; 1:40). Note that the 0.5 mL min-1 rate only applied 30 g NH4+ kg-1 at maximum, thus the sorption value at 40 g NH4+ kg-1 addition was not estimated. The fastest flow rate (4.5 mL min-1) resulted in sorption of only 7 and 9 g NH4+ kg-1 for zeolites A and D, respectively (Fig. 3), with addition of 40 g NH4+ kg-1. These results suggest that the removal of reaction products and maintaining a high concentration of sorptive is important for the reaction to proceed (i.e., three of the four flow rates sorbed an amount of NH4+ greater than or equal to the batch-shaken), but sufficient contact time is also needed for the reaction to occur (i.e., the fastest flow rate sorbed slightly less NH4+ than the batch even though the products were being removed). Interestingly, for the 1 mL min-1 rate, the total amount of NH4+ sorbed (Fig. 3) slightly exceeded the maximum NH4+ sorption based on measured CEC (maximum of 19.4 and 22.8 mg Fig. 3. Cumulative NH4+ sorbed onto zeolite expressed as a function of cumulative NH4+ added. added over 5 h in a flow-through cell at four different flow rates (4.5, 2.5, 1, NH4+ kg-1 for zeolites A and D, respectively, Synthetic effluent and 0.5 mL min−1). Zeolite A and D are shown in (a) and (b), respectively. based on measured CEC). However, note that the potential CEC for the zeolite minTable 3 presents the “breakpoint,” or the time in which the eral clinoptilolite is 220 cmolc kg-1 (i.e., 39.6 mg NH4+ kg-1; slope of the relationship between time and cumulative NH4+ Boettinger and Ming, 2002). Considering the percentage of sorbed (Fig. 4) significantly changed. Figure 4 and Table 3 clinoptilolite reported by the zeolite distributors (Table 2), this indicate that the fast flow-through application rates (4.5 and equates to a potential CEC of 187 and 176 cmolc kg-1 com2.5 mL min-1) significantly (P < 0.01) level off in NH4+ sorppared with the measured CEC of 108 and 127 cmolc kg-1 on tion between 76 and 142 min. On the other hand, the two zeolites A and D, respectively. Similarly, Cooney et al. (1999) slow flow-through application rates (0.5 and 1.0 mL min–1) found that K+-saturated clinoptilolite exchanged 100 cmolc sorbed NH4+ more steadily (i.e., linearly) without reaching NH4+ kg-1, while Weber et al. (1983), Kithome et al. (1998), an apparent plateau (Fig. 4), although the slope does slightly and Bolan et al. (2004) measured 140, 93, and 156 cmolc kg-1 decrease, indicating a decrease in sorption rate. As expected, CEC, respectively. This demonstrates two points: the tradibecause the slower flow rates sorbed NH4+ more steadily, the tional unbuffered salt extraction CEC method (Sumner and 1.0 and 0.5 mL min-1 rates displayed less decrease in slope after Miller, 1996) underestimated zeolite CEC, and it will only be the breakpoint (42–54% decrease in slope) compared with the possible to fully saturate the potential CEC if very high NH4+ 4.5 and 2.5 mL min-1 flow rates (81–100% decrease; Table 3). concentrations are used and/or reaction products are removed Comparing the 2.5 and 4.5 mL min-1 rates, the former not from the system (i.e., flow-through applications or use of many only sorbed more NH4+ after a given time period (Fig. 4) but “loading” and “stripping” steps for a batch test). Consider that also sorbed more efficiently when normalized based on NH4+ the CEC method utilized is essentially a “batch-shaken” experadded (Fig. 3). On the other hand, the sorption efficiency for iment. Boettinger and Ming (2002) reported that “actual” the two low rates (1.0 and 0.5 mL min-1) was the same (Fig. or measured zeolite CEC is ultimately affected by the ionic 3). These results suggest that differences in sorption curves strength and cation composition of the external solution. among different flow rates are probably a function of the rate
Table 3. Parameters from the results of the SAS “linear-plateau” model fitted to data shown in Figure 4 (P < 0.01). Parameter
Flow rate 4.5 mL min–1
2.5 mL min–1
BO† B1‡ Breakpoint§ B2¶
1.2 0.12 76 0.022
2.5 0.10 142 –0.005
BO B1 Breakpoint B2
3.5 0.085 93 0.011
3.2 0.15 106 0.007
1.0 mL min–1
0.5 mL min–1
1.1 0.11 126 0.055
0.53 0.066 154 0.033
0.96 0.12 124 0.055
0.33 0.069 137 0.04
Zeolite A
Zeolite D
† Intercept of the line prior to the breakpoint. ‡ Slope of the line prior to the breakpoint. § X value (time, min) in which the slope significantly changes. ¶ Slope of the line after the breakpoint.
of NH4+ load applied and the contact time for each flow rate. In other words, a sufficiently high contact time is necessary for the NH4+ sorption reaction to occur efficiently (relative to NH4+ added), and if this condition is met then the faster
NH4+ application rates (i.e., lower contact time) will result in a higher sorption rate (with respect to time). This is why the 1.0 mL min-1 flow rate sorbs NH4+ equal to the 0.5 mL min-1 flow rate based on NH4+ loading (Fig. 3) yet is able to sorb more NH4+ after a given time period (Fig. 4, Table 3). Nguyen and Tanner (1998) conducted a flow-through study on zeolites at two different flow rates using a synthetic effluent, a dairy effluent, and a swine effluent. Similar to our results, the authors found that the NH4+ breakthrough capacity increased by 29 to 57% with a decrease in flow rate from 1.59 to 0.047 mL min-1; these flow rates corresponded to a contact time of 17.8 h and 19.2 min, respectively.
Heat of Sorption and Contact Time
Fig. 4. Cumulative NH4+ sorbed onto zeolite expressed as a function of time. Synthetic effluent added over 5 h in a flow through cell at four different flow rates (4.5, 2.5, 1, and 0.5 mL min−1). Zeolite A and D are shown in (a) and (b), respectively.
The measured heat from NH4+ addition to zeolites A and D provided an indicator of the NH4+ sorption reaction (i.e., K+–NH4+ exchange). As reactions proceed, heat is either emitted (exothermic) or absorbed (endothermic), serving as an indicator of the degree to which a reaction has occurred over time. Figure 5 illustrates the heat pattern for addition of NH4+ to zeolites A and D at 25°C. Note the endothermic nature of the reaction occurring after NH4+ addition (negative values below the x axis indicate that the heater portion of the calorimeter was required to maintain a constant 25°C). This suggests that K is thermodynamically preferred over NH4+ on the zeolite, which agrees with cation selectivity studies conducted on zeolites (Ames, 1967; Al’tshuler and Shkurenko, 1992; Weatherley and Miladinovic, 2004; DeSutter and Pierzynski, 2005; Mumford et al., 2008). Kithome et al. (1999) concluded, based on calculated Gibbs free energy values for ion exchange onto zeolites previously saturated with either Ca2+, Mg2+, or K+, that NH4+ was not preferred. Using flow calorimetry, Rhue et al. (2002) showed that K+–Ca2+ exchange on an
Journal of Environmental Quality • Volume 39 • July–August 2010
Oxisol and K+–Na+ exchange on an Ultisol was endothermic; as expected, the reverse reactions were exothermic. Although K+ and NH4+ ions are very similar, the slightly lower hydration energy (335 vs. 339 kJ mol-1 for NH4+ and K+, respectively) and larger ionic radius (0.140 and 0.133 nm for NH4+ and K+, respectively) of NH4+ means that K+ will be slightly preferred over NH4+ with regard to ion exchange reactions on permanently charged minerals such as zeolites. In addition, the heat patterns provide useful information with regard to the timescale in which the exchange reaction proceeded. However, one cannot separate the rates of chemical and transport processes based on the results of this study. For example, Fig. 5 indicates that there is an initial fast reaction that occurs during the first 100 s after NH4+ addition; after that, the reaction slows down, as indicated by the decreasing slope. Fig. 5. Thermogram from titration of 250 mL of 0.7 M NH Cl into 0.1 g of zeolite A. 4 The decrease in reaction is probably due to diffu- Thermogram for zeolite D was identical except that the peak occurred at 408 s. sion of NH4+ into the micropores of the zeolite and accumulation of reaction products (i.e., K+). both of these flow rates exceeded the 100-s threshold for the After the peak, the continued slow climb occurs until about fast reactions, they sorbed NH4+ at the same efficiency based 3000 s (50 min). on NH4+ loading (Fig. 3). However, the 1.0 mL min-1 flow rate Cation exchange on zeolites is known to be controlled by removed more NH4+ after a given time period compared with diffusion (Boettinger and Ming, 2002; Kithome et al., 1998). the 0.5 mL min-1 flow rate (Fig. 4) simply because the former + Kithome et al. (1998) found that the NH4 sorption onto zeorate added more NH4+ per unit time; it was able to sorb this lite minerals conformed well to the parabolic diffusion model additional NH4+ because the contact time was sufficiently high and that either intraparticle diffusion or surface diffusion could enough (i.e., >100 s). These data suggest that flow-through be rate limiting (Crank, 1976). The authors described the NH4+ calorimetry can provide a rapid means of determining the optidiffusion process in five steps: (i) diffusion of NH4+ through the mum retention time for exchange reactions on zeolites. liquid, (ii) diffusion of NH4+ into the particle itself, (iii) chemical exchange between NH4+ and the exchangeable cation at the Summary and Implications exchange site, (iv) diffusion of the exchanged cation out of the When zeolites were contained in a mesh sack, suspended in a interior of the zeolite, and (v) diffusion of the displaced cation solution of swine effluent, and stirred (batch-stirred system), through the solution away from the zeolite mineral. For the NH4+ removal was lower than in the batch-shaken system flow-through system tested in the current study, steps i and v conducted at the same solid:solution ratio and equilibraare less relevant compared with batch systems due the solution tion times. Diffusion into the mesh sack reduced the rate of moving through the solid minerals. In addition, the constant NH4+ adsorption in the batch-stirred system compared with replenishment of solution NH4+ by flow-through increases the the shaken system where this diffusion mechanism was not “mass action” mechanism compared with a batch system. present. Furthermore, the flow-through analysis showed that Application of these data to NH4+ sorption systems suggests contact time is a critical consideration when optimizing NH4+ that approximately 100 s is required for the initial fast reacsorption to zeolites. The results suggest that in regards to praction to occur. This reaction is probably dominated by cation tical applications for removing NH4+ from solutions using zeoexchange on the zeolite surface. Recall that the 0.5, 1.0, 2.5, lites, a sufficiently high contact time (>100 s) and the ability to and 4.5 mL min-1 flow rates corresponded to contact times of remove reaction products while maintaining high NH4+ con330, 170, 70, and 30 s, respectively, and that the 2.5 and 4.5 centrations are required for achieving the maximum removal mL min-1 flow rates resulted in plateau in the adsorption curves efficiency (i.e., mass of NH4+ sorbed per mass NH4+ added). presented in Fig. 3, indicating a reduction in the number of Overall, the results of this study indicate implementation of effective exchange sites available for NH4+ adsorption. In cona flow-through contact time slightly higher than the “threshtrast, flow-through rates that provided contact times in excess old” time (i.e., 100 s) will optimize the rate and magnitude of 100 s (0.5 and 1.0 mL min-1) allowed for a greater proporof NH4+ adsorption on to zeolites. This threshold time can be tion of the NH4+ to adsorb (Fig. 3). Apparently, retention times rapidly determined using calorimetry. Increasing the contact that are less than the time required for the fast reaction to occur time beyond the “threshold” time will not drastically increase do not allow for complete diffusion of NH4+ and K+ within the NH4+ removal efficiency. However, it will reduce the rate of internal pore spaces of the zeolite, which limits the maximum NH4+ removal. Contact times that are shorter than the “threshamount of NH4+ adsorption. old” time will reduce both the NH4+ removal efficiency and The 100-s threshold contact time also explains the behavior sorption rate. The decision to use a batch vs. flow-through of the 0.5 and 1.0 mL min-1 rates. Because the contact time for
system and choice of flow-through rates will likely depend on the specific application, objectives, and cost for each system. For the swine industry, consider the following example: 1.6-ha lagoon with 6-m depth. In this application, the lower 4.8 m would not be treated due to maintaining anaerobic digestion at this depth. Assuming a flow-through system with a retention time of 170 s and effluent composition described in this study, zeolite A or D would remove at least 25 g NH4+ kg-1 zeolite. If 181 Mg (200 tons) of zeolite contained in this system was used to treat this effluent it could remove 4535 kg total NH4+ worth approximately $2955 based on the current value of fertilizer N in Oklahoma (NPK, 2010). The purchase of 181 Mg of zeolite would be a one-time cost of $12,000 at a price of approximately $66 per Mg (St. Cloud Zeolite, Truth or Consequences, NM) and could easily be contained in a 1.7 kL (4500 gal) tank, which costs less than $3000 (Tank Depot, Babylon, NY). Therefore, the recovered N could potentially offset the cost of zeolites after five to six cycles given that a recycling system using KCl solution is utilized, which would allow the N to be washed from the zeolites and then be reused in crop production. Additionally, future efforts to utilize zeolites to reduce NH3 emissions from swine production facilities may benefit from these findings.
References Al’tshuler, G.N., and G. Yu. Shkurenko. 1992. Thermodynamics of cation exchange on natural heulandite. Izv. Akad. Nauk. Ser. Khim. 3:754–756. Ames, L.L., Jr. 1967. Zeolite removal of ammonium ions from agricultural and other waste waters. p. 135–152. In Proc. Pacific Northwest Ind. Waste Conf., 13th, Pullman, WA. 6–7 Apr. 1967. Washington State Univ. Tech. Ext. Serv., Pullman. Bernal, M.P., J.M. Lopez-Real, and K.M. Scott. 1993. Application of natural zeolites for the reduction of ammonia emissions during the composting of organic wastes in a laboratory composting simulator. Bioresour. Technol. 43:35–39. Boettinger, J.L., and D.W. Ming. 2002. Zeolites. p. 585–610. In W.A. Dick and D.G. Schulze (ed.) Soil mineralogy with environmental applications. SSSA Book Ser. 7. SSSA, Madison, WI. Bolan, N.S., A.D. Adriano, and C. Wong. 2004. Nutrient removal from farm effluents. Bioresour. Technol. 94:251–260. Carpenter, S.R., N.E. Caraco, D.L. Corell, R.W. Howarth, A.N. Sharpley, and V.H. Smith. 1998. Nonpoint pollution of surface waters with phosphorus and nitrogen. Ecol. Appl. 8:559–568. Cooney, E.L., N.A. Booker, D.C. Shallcross, and G.W. Stevens. 1999. Ammonia removal from wastewaters using natural Australian zeolite: II. Pilotscale study using continuous packed column process. Sep. Sci. Technol. 34:2741–2760. Crank, J. 1976. The mathematics of diffusion. Oxford Univ. Press, New York. DeSutter, T.M., and G.M. Pierzynski. 2005. Evaluation of soils for use as liner materials: A soil chemistry approach. J. Environ. Qual. 34:951–962. Fernando, W.A.R.N., K. Xia, and C.W. Rice. 2005. Sorption and desorption of ammonium from liquid swine waste in soils. Soil Sci. Soc. Am. J. 69:1057–1065. Hamilton, D.W., W.G. Luce, and A.D. Heald. 1997. Production and characteristics of swine manure. Oklahoma State Univ. Extension Fact Sheet F-1735. Oklahoma Cooperative Extension Service, Stillwater. Huang, Z.T., and A.M. Petrovic. 1994. Clinoptilolite zeolite influence on ni-
trate leaching and nitrogen use efficiency in simulated sand based golf greens. J. Environ. Qual. 23:1190–1194. Kallay, N., V. Tomasic, S. Zalac, and L. Brecevic. 1997. Calorimetric investigation of kinetics of solid phase dissolution: Calcium carbonate dissolution in aqueous EDTA solution. J. Colloid Interface Sci. 188:68–74. Kithome, M., J.W. Paul, L.M. Lavkulich, and A.A. Bomke. 1998. Kinetics of ammonium adsorption and desorption by the natural zeolite clinoptilolite. Soil Sci. Soc. Am. J. 62:622–629. Kithome, M., J.W. Paul, L.M. Lavkulich, and A.A. Bomke. 1999. Thermodynamics of ammonium exchange by clinoptilolite. Commun. Soil Sci. Plant Anal. 30:895–907. LACHAT. 1994. QuickChem method 12-107-04-1-B. LACHAT Instrument, Milwaukee, WI. Matschonat, G., and E. Matzner. 1995. Quantification of ammonium sorption in acid forest soils by sorption isotherms. Plant Soil 168:95–101. Ming, D.W., and J.B. Dixon. 1987. Quantitative determination of clinopitolite in soils by a cation exchange capacity method. Clays Clay Miner. 35:463–468. Mumford, K.A., D.C. Shallcross, I. Snape, and G.W. Stevens. 2008. Application of a temperature-dependent semiempirical thermodynamic ionexchange model to a multicomponent natural zeolite system. Ind. Eng. Chem. Res. 47:8347–8354. Nelson, D.W., and L.E. Sommers. 1996. Total carbon, organic carbon, and organic matter. p. 961–1010. In D.L. Sparks (ed.) Methods of soil analysis. Part 3. Chemical methods. SSSA Book Ser: 5. SSSA and ASA, Madison, WI. Nguyen, M.L., and C.C. Tanner. 1998. Ammonium removal from wastewaters using natural New Zealand zeolites. N. Z. J. Agric. Res. 41:427–446. NPK. 2010. Department of plant and soil science. Available at http://www. npk.okstate.edu/ (verified 20 Apr. 2010). Panetta, D.M., W.P. Powers, and J.C. Lorimor. 2005. Management strategy impacts on ammonia volatilization from swine manure. J. Environ. Qual. 34:1119–1130. Rhue, R.D., C. Appel, and N. Kabengi. 2002. Measuring surface chemical properties of soil using flow calorimetry. Soil Sci. 167:782–790. Rozic, M., S. Cerjan-Stefanovic, S. Kuraijica, V. Vancina, and E. Hodzic. 2000. Ammoniacal nitrogen removal from water by treatment with clays and zeolites. Water Res. 34:3675–3681. SAS Institute. 2003. SAS user’s guide: Statistics. SAS Inst., Cary, NC. Seol, Y.K., and L.S. Lee. 2000. Effect of dissolved organic matter in treated effluents on sorption of atrazine and prometryn by soils. Soil Sci. Soc. Am. J. 64:1976–1983. Sumner, M.E., and W.P. Miller. 1996. Cation exchange capacity and exchange coefficients. p. 1201–1231. In D.L. Sparks (ed.) Methods of soil analysis. Part 3. Chemical methods. SSSA Book Ser. 5. SSSA, Madison, WI. Weatherley, L.R., and N.D. Miladinovic. 2004. Comparison of the ion exchange uptake of ammonium ion onto New Zealand clinoptilolite and mordentite. Water Res. 38:4305–4312. Weber, M.A., K.A. Barbarick, and D.G. Westfall. 1983. Ammonium adsorption by a zeolite in a static and a dynamic system. J. Environ. Qual. 12:549–552. Wu, J., D.L. Nofziger, J. Warren, and J. Hattey. 2003. Modeling ammonia volatilization from surface-applied swine effluent. Soil Sci. Soc. Am. J. 67:1–11. Zaman, M., M.L. Nguyen, F. Matheson, J.D. Blennerhassett, and B.F. Quin. 2007. Can soil amendments (zeolite or lime) shift the balance between nitrous oxide and dinitrogen emissions from pasture and wetland soils receiving urine or urea-N? Aust. J. Soil Res. 45:543–553. Zaman, M., M.L. Nguyen, and S. Saggar. 2008. N2O and N2 emissions from pasture and wetland soils with and without amendments of nitrate, lime, and zeolite under laboratory condition. Aust. J. Soil Res. 46:526–534. Zhao, Y., T. Gao, S. Jiang, and D. Cao. 2004. Ammonium removal by modified zeolite from municipal wastewater. J. Environ. Sci. (China) 16:1001–1004.
Journal of Environmental Quality • Volume 39 • July–August 2010
