- Email: [email protected]
An evaluation of pretreated natural zeolites for ammonium removal
Kater Res,'arch Vol. 14. pp. 161 to 168 © Pergamon Press Ltd 1980. Printed in Gfcal Britain
{X)43-1354 80/0201-0161502.00/0
AN EVALUATION OF PRETREATED NATURAL ZEOLITES FOR A M M O N I U M REMOVAL J. R. KLIEVE E n v i r o n m e n t a l R e s e a r c h E n g i n e e r , M o n s a n t o R e s e a r c h Corp., D a y t o n , O H 45418, U.S.A.
and M. J. SEMMENS Associate Professor-Environmental Engineering, Department of Civil and Mineral Engineering, University of Minnesota, Minneapolis, MN 55455, U.S.A. (Received 28 June 1979)
Abstraet--Clinoptilolite has been widely studied for ammonium removal in the past 2 yr. However, many investigators have reported variations in the measured capacities of samples of clinoptilolite. These studies and the factors believed to influence measured zeolite capacity are reviewed. In addition no studies to evaluate other natural zeolites for ammonium removal have been reported. In this study samples of clinoptilolite, erionite, mordenite and phillipsite provided by the Anaconda Company were evaluated for ammonium removal from wastewaters. In addition, samples of clinoptilolite were pretreated in various ways to determine whether an improvement in ammonium removal performance could be realized. Total exchange capacities, capacities for ammonium removal from a synthetic waste, packed bed densities and crushing strengths were measured. Phillipsite was found to have almost twice the weight capacity for ammonium removal from synthetic waste compared to that of clinoptilolite. The volumetric capacity was 26% better than that of clinoptilolite. However, the phillipsite sample was extremely friable and it could not be used for water treatment unless it was strengthened with a binder. Pretreatment of clinoptilolite with NaOH, HNO3 and steam did little to improve the zeolite's performance, However, heat pretreatment (600°C for 1 h) improved the zeolite's selectivity for ammonium significantly. Ammonium removal capacities were increased by approximately 177/~,for heat treated zeolite samples although the total exchange capacity of the zeolite was reduced somewhat.
INTRODUCTION
Nitrogen removal from wastewaters by selective ion exchange has been studied widely in the last I0 yr. Clinoptilolite, a naturally occurring zeolite, has been found selective for ammonium ions in the presence of other cation concentrations commonly found in wastewaters. Ammonium removal by clinoptilolite is cost competitive with other total nitrogen removal processes. The clinoptilolite is normally crushed and graded when it is mined before being used as an ion exchange medium. However, the capacity of clinoptilolite is influenced significantly by chemical and physical treatments. In practice facilities using untreated clinoptilolite may observe an improvement in performance following several service-regeneration cycles. Figure 1 shows breakthrough curves obtained by the authors for ammonium removal by clinoptilolite in which curve I was obtained with freshly graded untreated zeolite and curve II was obtained after several serviceregeneration cycles. The influence of pretreatments on zeolite capacity and selectivity is poorly documented, and where information exists, care must be exercised in comparing the zeolite capacities and pretreatment effects reported. The capacity of a zeolite varies with the source of the zeolite, the location within a particular deposit and the capacity measurement technique W.R. 14/2
I>
employed. Past studies evaluating the capacity of clinoptilolite samples that had received slight pretreatment were summarized by Koon & Kaufman (1971) as shown in Table 1. It can be seen that slight acid pretreatment as employed by Ames (1962, 1964) did nothing to improve the clinoptiiolite's capacity. Acid would be expected to dissolve alkali impurities which may block pores and/or add weight to the zeolite. Murphy et al. (1976) compared various pretreated clinoptilolite samples for ammonium removal from municipal wastewater. Hector clinoptilolite was contacted in a batch pretreatment process with the solutions of(l)acid, (2) alkali, (3) alkali followed by NaCI and (4) acid followed by NaOH and then NaCI. Capacities were measured using a sequential batch equilibration (2 h) technique. The authors concluded that pretreatment 4 resulted in the greatest increase in exchange capacity, 20~ over samples lacking pretreatment. The other pretreatments resulted in an increase in the exchange capacity of 5~o or less. Additional studies indicated that the influence of a pretreatment on zeolite capacity varied widely depending on the source of the clinoptilolite. Seyforth (1975) investigated the influence of pretreatment on the capacity of clinoptilolite (Buckhorn, New Mexico) for ammonium ions. To ensure meaningful capacity values were measured, a zeolite sample
161
162
J.R. KLttivt and M. J. SEMMENS I0 Curve I - Unconditioned zeolite Curve 2 - Conditioned zeoite
08
~
. , , ~ o °Curve 2 e-I
06
04
"/
02
0.0
0
t00
2 O0
300
400
Bed volumestreated
Fig. I. The impact of conditioning on ammonium breakthrough curves for clinoptilolite (influent NH,~--N = 20 mg/I, 20 Bv/hr).
was conditioned by several alternate treatments with conditioned and reconditioned zeolite capacities in 0.4 N NH,CI and 1 N NaCI. Portions of this sample Table 2. Since the reconditioned zeolite was exposed were given additional treatments. One portion was to an additional two conditioning cycles, one might reconditioned repeating the conditioning procedure expect the first capacity test with this sample to be but an alkaline (pH 10) 1N NaCI regenerant was approximately equal to the third measured capacity employed. The other portion was exposed to excess of the conditioned zeolite. There is, however, a signifi10% HNO3 for 1 h and then 1 N NaCI to convert the cant difference in these values, and the first capacity zeolite to the Na + form, All zeolite samples were measurement of the conditioned and reconditioned samples are closer in value. dried before use for convenience. In summary, therefore, the capacity of natural zeoThe average capacities of the several different preparations of clinoptilolite were not the same as shown lite is influenced significantly by the pretreatments the in Table 2. The capacity of the conditioned zeolite, zeolite has received and the influence of various pre1.69 + 0.08 meq/g dry zeolite, was smaller than those treatments is poorly documented. In addition, zeolites other than clinoptiiolite have not been evaluated for of the reconditioned and acid-washed zeolite, 1.82 + 0.23 and 1.86 + 0.05 meodg dry zeolite, re- removing ammonium from wastewater by ion spectively. Although the latter two preparations gave exchange. Hayhurst (1976) evaluated mordenite, philapproximately the same average capacity, the varia- lipsite and clinoptilolite for removing NH3 from gas bility in the value of the reconditioned zeolite was streams. Phillipsite was demonstrated to be best in much greater than that of the acid-washed zeolite. removing NH3 from gas. This paper presents the results of some laboratory Reconditioning with a salt solution adjusted to pH 10.0 should have removed acidic impurities from the studies to evaluate phillipsite, mordenite, erionite and zeolite. Similarly, acid-washing should have removed pretreated clinoptilolites for ammonium removal from wastewater. Ammonium capacities were any alkali impurities. When capacity tests on the same zeolite sample measured in two ways: (1) using a 100mg/l solution were repeated following regeneration with 1 N NaCI, of NH,~--N as a feed solution in column exhaustion the zeolite capacities were observed to increase. After tests and (2) using a synthetic sewage containing repeating the test three times the capacity appeared to 14mg/l N H 2 - - N as a feed solution in column exhaustion tests. Packed bed densities were deterlevel off at 2.24 meq/g dry zeolite. Jorgensen et al. (1976) also noted that the capacity mined to convert observed capacities on a weight of of a sample of clinoptilolite increased with the zeolite basis to capacities on a volume of zeolite basis. number of regenerations. Their regeneration process A crushing test was used to assess physical differences consisted of treatment with 4% NaOH for 30 min, in the zeolites. followed by a rinse with 11 of deionized water for 40 min. After repeating this procedure four times, the measured capacity was found to be constant. They LABORATORY TECHNIQUES AND EQUIPMENT concluded that the sodium hydroxide "activates" the clinoptilolite. The results of Seyforth's study suggest Grading and pretreating the zeolites Twenty-five pound samples of erionite, mordenite, that it was not necessarily the sodium hydroxide which was responsible for the increased capacity. It and phillipsite were provided by Anaconda Company, may, however, be related to the number of exchange Denver, Colorado for use in this study. Clinoptilolite cycles that the zeolite is exposed to after being oven from Buckhorn, New Mexico was provided by dried at 103°C. This may be seen by comparing the Double Eagle Mining Co. (Casper, Wyoming) and
Pretreated natural zeolites
"F-, .~ ~
~g
2~
0
"" ,~
163
z'~z~ o ~
.~:-~
~:'~'.="-~,,~.o~oo~. •- + 0
.~.~
~;
.~'~o-0
~o
= ~
~
o
0 0
e-
[-
.~__ .o .o ._.
~. ~.=_
_q
Z
0 'n ~
0 '."
Z
0 0 ,m
l-
g
~
~,~ o~
Z
X
X
X
X
o
o
~
o
8
~o +~
Z
o
~8
n-
8
e~
o
o ÷
F..
o
..0 n~
,.o'~
"~ ~
+~
+~
.+_+~
Zo~
Z
Z
o
0 0
g.
0 ..~,
0
e~
0
g
g u~
o ..d 0
0
164
J.R. KLH!Vl and M. J. SEMMENS Table 2. Results of the capacity tests conducted by Seyforth (16) Type of clinoptilolite sample
Number of samples tested
Capacity meq g t
Sample test number
5 3 2 6 2 4*
1.69 + 0.08 2.11 +_ 0.10 2.23 + 0.01 1.82 + 0.23 2.25 + 0.07 1.85 + 0.05
1 2 3 I 2 1
Conditioned Conditioned Conditioned Reconditioned Reconditioned Acid washed
* Capacities measured by both ammonium and sodium elution studies. graded with U.S. standard mesh sieves. The 20 × 30 mesh fractions were retained for use in this study. Zeolite samples were pretreated in different ways. Solutions were pumped from a feed reservoir upflow through a 2.5 cm acrylic column containing approximately 80 g of a zeolite sample. A nylon screen supported the zeolite in the column. Table 3 presents the various pretreatment conditions employed to prepare the zeolite samples. All samples were treated with approximately 41 of l N NaC1 over a 2 h period. Some samples were given two treatments. Following NaCI treatment zeolite samples were washed with deionized water. Nitric acid treatment was accomplished by contacting a zeolite sample with 21 of 1 N HNO3 over a l h period. Samples 5 and 5B were contacted with l N N a O H treatments of 21 in 70 min and 31 in 1 l0 rain, respectively. The heat treated zeolite samples were heated to 60ffC for 1 h, cooled overnight and rewetted. Some samples were autoclaved at 30psi g 134°C for 1 h.
tion of ammonium chloride (100mg/l NH~--N) in deionized water. The feed solution was supplied to the column from an elevated reservoir at a flow rate of 20 bed vois/h (10 ml/min), 1 1 samples of the column effluent were collected and analyzed for ammonium concentration. Samples of the feed were analyzed for ammonium concentration each time an effluent sample was collected to ensure there was no change in the feed ammonium concentration during the test period. When the effluent ammonium concentration was the same as that of the feed (within analytical error), the zeolite was flushed with deionized water, removed from the column, dried and weighed. Total operating capacity measurements. These measurements were made using exactly the same procedure employed for total capacity measurements except that the feed solution had the composition presented in Table 4. This solution composition was identical to that used by Breck (1973) and contained no organics or colloids.
Testing the zeolite,s Measurements were made of the total ammonium capacities of the pretreated zeolites. In addition, their capacities for ammonium in the presence of competing ions were measured to determine any improvement in selectivity. Total capacity measurements. A 50 ml burette containing 30ml of pretreated zeolite supported on a glass wool plug was contacted downflow with a solu-
Analytical techniques Ammonium measurements were made using an Orion Model 95-10 ammonia selective electrode connected to an Orion Model 701 digital pH/mV meter (Orion, Cambridge, Mass.). The electrode tended to drift and standard ammonia samples were analyzed between every analysis of an unknown sample as a check on the calibration curve. Duplicates of the an-
Table 3. Pretreatment scheme for clinoptilolite samples tested
Sample Clinoptilolite I Clinoptilolite I B Clinoptilolite 2 Clinoptilolite 2B Clinoptilolite 3 Clinoptilolite 3B Clinoptilolite 4 Clinoptilolite 5 Clinoptitolite 5B Phillipsite Mordenite Irionite
HNO3 + rinse
NaOH + rinse
Second NaCI + rinse
First NaCI
Backwash
X
X
X
X
X
×
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
×
X
X
X
X
X
X
Heat treatment Autoclaving
X X X
165
Pretreated natural zeolites Table 4. Composition of the synthetic waste water (14) Element
Concentration
Compound used
NH2 Na + K+ Mg + + Ca + + pH
14 mg/l 58 mg/l 12 mg/l
(NH4): SO4 NaCI, NaHCO3 KC1 MgSO4 CaCI2.2H20 Adjusted with proper ratio of NaHCO3 to NaCI
8 mg/I
34 mg/l 7.45
Table 5. Capacity of zeolites using 100 nag NH~'--N/I solution as the feed solution Capacity (meq NH+--N/g) Individual tests Mean
Zeolite Clinoptilolite 1 Clinoptilolite 2 Clinoptilolite 3 Clinoptilolite 5 Phillipsite Mordenite Erionite
1.70, 1.51, 1.62, 1.85, 2.73, 1.76, 0.30
known samples and occasionally triplicates of the unknown samples were run when the drift was excessive. To weigh the zeolites in a reproducible way wet zeolite samples were dried in a 100°C oven overnight and then placed in a desiccator over a saturated solution of NaC1. The desiccator was placed in a constant temperature (37°C) oven to maintain a constant humidity (75%) environment (15). The samples were held in the desiccator for 2-3 days to assure equilibrium with the humidity in the desiccator.
Zeolite packed bed densities Since the specific gravity and packing characteristics of the zeolites vary, packed bed densities were measured in order that volumetric capacities could be determined. Volumetric capacities are important for sizing ion exchange equipment. A 25 ml graduated cylinder filled partially with water was filled to 25 ml with wet zeolite. The cylinder was rapped on the lab counter until the zeolite volume showed no further reduction~ Final volume was recorded. The samples were removed from the cylinder, dried and weighed.
Zeolite strenyth To compare the physical strength of the zeolite samples a crushing test developed by Hayhurst (1976) was used. Zeolites employed in column operations are prone to physical attrition during backwashing and regeneration cycles. It is desirable, therefore, to compare the physical strength of the zeolites in order to determine their suitability for column operation. A single zeolite grain was placed on the weighing pan of a top loading direct reading Mettler Balance (Princeton, New Jersey). The balance was positioned beneath a drill press fitted with a 1/4 in steel rod. The
1.61, 1.61, 1.50, 1.73, 2.55 1.83
1.60, 1.76, 1:57, 1.75,
1.61, 1.64 1.67, 1.65 1.57 1.75
1.63 1.64 1.57 1.77 2.64 1.80 0.30
rod was slowly lowered on to zeolite particle and pressure was applied gradually until the grain fractured. The weight required to crush the particle was recorded as a measure of crushing strength. The test was run on 20 grains of each zeolite sample to obtain a significant mean value for comparative purposes.
X-ray diffraction analyses Standard X-ray diffraction patterns were obtained for powdered samples of untreated and pretreated zeolites using a Phillips-Norelco Diffractometer. The instt'ument employed CuK ~ radiation at 1° 20/rain. RESULTS The results of the capacity tests are presented in Tables 5 and 6. The test techniques ensured excellent reproducibility as indicated by the standard deviation value§. Clinoptilolite samples 1, 2 and 4 all had an average total ammonium capacity of 1.63-1.64 meq/g. Sample 3 treated clinoptilolite had a somewhat lower capacity of 1.57 meq/g and sample 5, the sample that received a treatment of sodium hydroxide, had the highest observed capacity of 1.77 meq/g. Table 6. Capacity of zeolites using synthetic sewage as the feed solution Zeolite
Capacity (meq NH2--N/g) Individual tests Mean
Clinoptilolite 1 Clinoptilolite 2 Clinoptilolite 3 Clinoptilolite 5 Phillipsite Mordenite
0.499, 0:502, 0.498 0.500 0.530, 0.514, 0.526 "0.523 0.590, 0.594 0.592 0:500, 0.514 0.507 0.970, 1.031 1.001 0.439, 0.444 0.442
166
J.R. KLIEVI~and M. J. S~MM~:NS
Table 7. Packed bed densities of 20 × 30 mesh zeolite in Na ÷ form
batch of clinoptilolite. The reason for this difference in measured capacities may stem from the different techniques employed for measuring the weight of zeoPacked bed density lite. Seyforth dried the zeolite at 105°C, cooled it in a Zeolite g dry zeolite/submerged ml desiccator and weighed the dry zeolite. In this study it Clinoptilolite 1 0.848 was observed, however, that this technique could Clinoptilolite 2 0.844 cause significant variation in the measured weight of a Clinoptilolite 3 0.853 zeolite sample. Natural zeolites are extremely effective Clinoptilolite 4 0.849 dessicants themselves and the dried zeolite samples Clinoptilolite 5 0.857 Mordenite 0.699 compete with chemical dessicants for the available Phillipsite 0.534 water. Thus, if the dessicant used is partially Erionite 0.503 exhausted, the zeolite may extract some of the available water and as a result weigh heavier than when Mordenite had a similar capacity to clinoptilolite fresh dessicant is used. The preferred equilibration but erionite had a very much lower total exchange with a constant, high humidity atmosphere gives capacity of only 0.3 meq/g. Phillipsite, with a capacity more reproducible results but the absorbed water tends to make the weight capacity appear smaller. of 2.64 meq/g was clearly superior to clinoptilolite. Tests on the ammonium capacities measured folAs noted above, differences between the clinoptilolowing exposure to synthetic sewage indicated that lite sample studied and other zeolites preclude direct the heat treated clinoptilolite (sample 3) was most comparison of these results with those reported by selective for ammonium. The other clinoptilolite other investigators, but the values obtained (Table 5~ samples showed that the other pretreatment tech- do appear similar to those reported by others (Table 1). niques had little impact on ammonium selectivity. In this study phillipsite was observed to have a Phillipsite was equally selective for ammonium and weight capacity that was 70°,0 greater than that of its higher total capacity resulted in an effective capa- similarly treated clinoptilolite. Tests conducted on the city that was about double that of clinoptilolite on a selective removal of ammonium from a synthetic sewweight basis. age containing competing ions indicated that the philThe packed bed densities of the zeolite samples are lipsite was more selective for ammonium than clinoppresented in Table 7. The type of pretreatment had no tilolite and removed twice as much as ammonium on effect on the packed density of clinoptilolite which a dry weight basis. However, the low packed bed denremained at a value of approximately 0.85g dry t sity of the phillipsite compared to that of clinoptiiolite zeolite/ml. The phillipsite was considerably lighter, reduced the performance of phillipsite on a voluhaving a packed density of only 0.53 g dry zeolite/ml. metric basis. Its volumetric capacity was only 26~, The crushing tests indicated that samples I and 3 of better than similarly treated clinoptilolite for clinoptilolite had the highest measured crushing ammonium removal from synthetic sewage. Phillipsite has two disadvantages that must instrengths. The results are depicted in Table 8. The other pretreatments for clinoptilolite reduced the fluence its use in water treatment, however. Firstly, strength of the zeolite. Phillipsite was observed to be this zeolite is structurally very weak and breaks easily. extremely friable and yielded a very low crushing Even in limited laboratory studies physical attristrength. The standard deviations were poor for this tion of the phillipsite was significant and the fines test and the difference between samples 1 and 3 of produced created large headlosses in the column clinoptilolite reported in Table 8 cannot be con- studies. Secondly, phillipsite is considerably more expensive than clinoptiiolite as shown in Table 9. sidered significant. Possibly the attrition problem could be overcome by mixing powdered phillipsite with a porous binder DISCUSSION OF RESULTS but as can be seen the costs of the zeolite would not The observed capacities of the pretreated clinopti- be competitive with clinoptilolite. However, if lolite samples tested in this study appear a little lower through the use of a binder the packed bed density of than those reported by Seyforth (1975) for the same the zeolite could be increased without reducing the
Table 8. Results of crushing test Zeolite Clinoptiiolite 1 Clinoptilolite 3 Phillipsite Mordenite Erionite
Number of tests
Mean crushing strength (g)
Standard deviation
20 20 20 20 20
366 329 83 129 105
114 lit 63 100 51
G
Pretreated natural zeolites Table 9. Approximate cost of zeolites in bulk quantities (22) (FOB) Zeolite Clinoptilolite Phillipsite Mordenite Erionite
S/ton 200-210 40(0450 300 400
exchange capacity significantly, then this may prove feasible. Analysis of the volumetric capacities of the clinoptilolite indicated that the heat treated sample was the best for ammonium removal from synthetic sewage. The heat treated clinoptilolite had approximately 17% more capacity than clinoptilolite that had been exposed only to NaCI. This pretreatment improved the selectivity of the zeolite for NH2 in exchange with competing cations and its measured capacity approached that of phillipsite. The heat treatment caused the clinoptilolite to turn dark brown, suggesting that iron oxidation had occurred within the zeolite structure. Why this should improve the zeolite's selectivity for NH,~ is not clear. Studies conducted by Anaconda have shown that the iron content of clinoptilolite is not exchangeable nor is it leached by applied EDTA (personal communication, 1978). The X-ray diffraction studies on the variously treated clinoptilolite samples unfortunately provided no evidence of any structural changes. Indeed, all the X-ray diffraction scans on the pretreated samples were identical. Further studies are needed to characterize the influence of heat on the zeolite. In this study a fixed time at a fixed temperature was examined. The time and temperature of pretreatment may be important. In addition, the technique of rehydration may be influential in affecting zeolite performance. During the course of pretreatment the actual weight of a zeolite sample may change. Thus, capacities reported per unit weight of the pretreated zeolite may be misleading. For example, if pretreatment reduces the weight of a zeolite sample without influencing capacity, then naturally the exchange capacity/unit weight will appear higher. Yet no great advantage may be realized from such a pretreatment. In this study weight changes were recorded for different pretreatments. No significant weight changes were recorded apart from the following: 1. Heat treatment caused a 4% weight loss. Since the total exchange capacity also declined by 4% following heat treatment, it would appear that some exchange sites are lost during heat pretreatment. 2. NaOH pretreatment caused a slight weight gain, 2%. The improved total exchange capacity of the base pretreated zeolite must therefore stem from the exposure of more exchange sites. 3. Acid pretreatment caused a 9% weight loss and no impact on the exchange capacity which suggests
167
that zeolite structure itself may be broken down by the strongly acidic conditions employed in pretreatment. The ion exchange process is a diffusion controlled process and the kinetics of exchange may be influential in the choice of a zeolite where capacity and selectivity are more or less equal. Some preliminary data was collected from the column studies on the pretreated clinoptilolite samples that suggested that the kinetics of exchange were not significantly influenced by pretreatment. Additionally, the kinetics of exchange with phillipsite were comparable with clinoptilolite.
CONCLUSIONS several pretreated clinoptilolites and other natural zeolites have been evaluated for ammonium removal from wastewater. The following conclusions may be drawn from this study: 1. The weight of a natural zeolite is influenced significantly by the relative humidity. It was therefore necessary to weight zeolites equilibrated with a constant relative humidity in order to obtain accurate and reproducible capacity data. 2. By heat treating clinoptilolite, the zeolite's capacity for NH2 was not significantly affected but the zeolite capacity for N H 2 - - N in the presence of competing cations was observed to increase. The selectivity of the zeolite for ammonium was thus enhanced. 3. By treating the clinoptilolite with NaOH, capacity for N H 2 - - N using a pure solution of NH4CI as feed was slightly improved, but the zeolite's selective capacity for N H 2 - - N in the presence of competing cations was not affected. 4. Acid and steam treatments did not influence clinoptilolite's capacity for N H 2 - - N . 5. Erionite, mordenite and phillipsite samples (Anaconda) were also tested although no pretreatments were applied to these other than conversion to the sodium form by 1 N NaCI. The phiihpsite sample was very effective in ammonium removal but erionite and mordenite'behaved poorly. The phillipsite proved to be better than even pretreated clinoptilolite samples in both weight and volumetric capacities. Unfortunately, phillipsite.is structurally weak and breaks down easily to produce fines. This will limit application of this zeolite in water treatment. If the strength of the zeolite could be improved through the use of a binder, it may find a wider application.
REFERENCES
Ames L. L. (1962) Effect of base cation on the cesium kinetics of clinoptilolite. Am. Min. 47, 1310-1316. Ames L. L. (1964) Some zeolite equilibria with alkali metal cations. Am. Min. 49, 127-145.
168
J.R. KLniVl! and M. J. SEMMENS
Ames L. L. (1967) Zeolitic removal of ammonium ions from agricultural and other wastewaters. 13th Pacific Northwest Industrial Waste Conference Proceedings, Washington State University, Pullman, Washington. Barrer R. M., Papadopoulos R. & Rees L. V. C. (1967) Exchange of sodium in clinoptilolite by organic cations. J. lnor 9. Nucl. Chem. 29, 2047-2065. Breck D. W. (1973) U.S. Patent 3,723,308. Dean J. A. (19731 Lange's Handbook qf Chemistry. l lth Edition, McGraw-Hill, p. 10-79. Deffeyes (1959) Anl. Min. 44, 501. Frysinger G. R. (1962) Cesium-sodium ion exchange on clinoptilolite. Nature 194, 351 353. Harris & Brindly (1954) Am. Min. 39, 819-824. Hayhurst D. T. (1976) Potential use of natural zeolites for ammonia removal during coal gasification, presented at Zeolite 1967 Cot!ference, Tucson, Arizona. Hoss & Roy (1960) Beitr. Minerl. Petro,q. 7, 389. Howery D. G. & Thomas H. C. (1965) Ion Exchange on the mineral clinoptilolite. J. Phys. Chem. 69(2), 531 537. Jorgensen S. E. (1976) Ammonia removal by use of clinoptilolite. Water Res. 10, 213-224. Mumpton F. A. (1960) Clinoptilolite redefined. Am. Min. 45, 351-369.
Murphy C. B., Hrycyk O. & Gleason W. T. (19761 Natural zeolites: novel uses and regeneration in wastewater treatment, presented at Zeolite 1976 ConiC,fence, Tucson, Arizona. Obretenov Z., Dimitrove D., Todorova V. & Toncheva E. 11976) Production of absorbents with molecular sieve action from native bulgarian clinoptilolite. Wiss. Z. Tech. Hachsch. "'Carl Schorlemmer'~ Lcuna-Mersehur¢4, 18{4), 569-572. Personal communication with Mr. Dennis Leonard, Anaconda Company, 22 August 1977. Personal communication with Mr. Dennis Leonard, Anaconda Company, December 2, 1977. Personal communication with Dr. Harold Vincent, Anaconda Company, 2 December 1977. Personal communication with Dr. Harold Vincent, Anaconda Company, 30 August 1978. Seyforth M. C. (19751 The Selectivity c~f Clinoptilolite Jor Heary Metal Cations, M.S. Thesis, University of lllinois, Urbana. Weast R. C. (1974)(Ed.) Handbook oJ Chenlistry and Physics, 55th Edition. The Chemical Rubber Company Press, p. 10-79.
{X)43-1354 80/0201-0161502.00/0
AN EVALUATION OF PRETREATED NATURAL ZEOLITES FOR A M M O N I U M REMOVAL J. R. KLIEVE E n v i r o n m e n t a l R e s e a r c h E n g i n e e r , M o n s a n t o R e s e a r c h Corp., D a y t o n , O H 45418, U.S.A.
and M. J. SEMMENS Associate Professor-Environmental Engineering, Department of Civil and Mineral Engineering, University of Minnesota, Minneapolis, MN 55455, U.S.A. (Received 28 June 1979)
Abstraet--Clinoptilolite has been widely studied for ammonium removal in the past 2 yr. However, many investigators have reported variations in the measured capacities of samples of clinoptilolite. These studies and the factors believed to influence measured zeolite capacity are reviewed. In addition no studies to evaluate other natural zeolites for ammonium removal have been reported. In this study samples of clinoptilolite, erionite, mordenite and phillipsite provided by the Anaconda Company were evaluated for ammonium removal from wastewaters. In addition, samples of clinoptilolite were pretreated in various ways to determine whether an improvement in ammonium removal performance could be realized. Total exchange capacities, capacities for ammonium removal from a synthetic waste, packed bed densities and crushing strengths were measured. Phillipsite was found to have almost twice the weight capacity for ammonium removal from synthetic waste compared to that of clinoptilolite. The volumetric capacity was 26% better than that of clinoptilolite. However, the phillipsite sample was extremely friable and it could not be used for water treatment unless it was strengthened with a binder. Pretreatment of clinoptilolite with NaOH, HNO3 and steam did little to improve the zeolite's performance, However, heat pretreatment (600°C for 1 h) improved the zeolite's selectivity for ammonium significantly. Ammonium removal capacities were increased by approximately 177/~,for heat treated zeolite samples although the total exchange capacity of the zeolite was reduced somewhat.
INTRODUCTION
Nitrogen removal from wastewaters by selective ion exchange has been studied widely in the last I0 yr. Clinoptilolite, a naturally occurring zeolite, has been found selective for ammonium ions in the presence of other cation concentrations commonly found in wastewaters. Ammonium removal by clinoptilolite is cost competitive with other total nitrogen removal processes. The clinoptilolite is normally crushed and graded when it is mined before being used as an ion exchange medium. However, the capacity of clinoptilolite is influenced significantly by chemical and physical treatments. In practice facilities using untreated clinoptilolite may observe an improvement in performance following several service-regeneration cycles. Figure 1 shows breakthrough curves obtained by the authors for ammonium removal by clinoptilolite in which curve I was obtained with freshly graded untreated zeolite and curve II was obtained after several serviceregeneration cycles. The influence of pretreatments on zeolite capacity and selectivity is poorly documented, and where information exists, care must be exercised in comparing the zeolite capacities and pretreatment effects reported. The capacity of a zeolite varies with the source of the zeolite, the location within a particular deposit and the capacity measurement technique W.R. 14/2
I>
employed. Past studies evaluating the capacity of clinoptilolite samples that had received slight pretreatment were summarized by Koon & Kaufman (1971) as shown in Table 1. It can be seen that slight acid pretreatment as employed by Ames (1962, 1964) did nothing to improve the clinoptiiolite's capacity. Acid would be expected to dissolve alkali impurities which may block pores and/or add weight to the zeolite. Murphy et al. (1976) compared various pretreated clinoptilolite samples for ammonium removal from municipal wastewater. Hector clinoptilolite was contacted in a batch pretreatment process with the solutions of(l)acid, (2) alkali, (3) alkali followed by NaCI and (4) acid followed by NaOH and then NaCI. Capacities were measured using a sequential batch equilibration (2 h) technique. The authors concluded that pretreatment 4 resulted in the greatest increase in exchange capacity, 20~ over samples lacking pretreatment. The other pretreatments resulted in an increase in the exchange capacity of 5~o or less. Additional studies indicated that the influence of a pretreatment on zeolite capacity varied widely depending on the source of the clinoptilolite. Seyforth (1975) investigated the influence of pretreatment on the capacity of clinoptilolite (Buckhorn, New Mexico) for ammonium ions. To ensure meaningful capacity values were measured, a zeolite sample
161
162
J.R. KLttivt and M. J. SEMMENS I0 Curve I - Unconditioned zeolite Curve 2 - Conditioned zeoite
08
~
. , , ~ o °Curve 2 e-I
06
04
"/
02
0.0
0
t00
2 O0
300
400
Bed volumestreated
Fig. I. The impact of conditioning on ammonium breakthrough curves for clinoptilolite (influent NH,~--N = 20 mg/I, 20 Bv/hr).
was conditioned by several alternate treatments with conditioned and reconditioned zeolite capacities in 0.4 N NH,CI and 1 N NaCI. Portions of this sample Table 2. Since the reconditioned zeolite was exposed were given additional treatments. One portion was to an additional two conditioning cycles, one might reconditioned repeating the conditioning procedure expect the first capacity test with this sample to be but an alkaline (pH 10) 1N NaCI regenerant was approximately equal to the third measured capacity employed. The other portion was exposed to excess of the conditioned zeolite. There is, however, a signifi10% HNO3 for 1 h and then 1 N NaCI to convert the cant difference in these values, and the first capacity zeolite to the Na + form, All zeolite samples were measurement of the conditioned and reconditioned samples are closer in value. dried before use for convenience. In summary, therefore, the capacity of natural zeoThe average capacities of the several different preparations of clinoptilolite were not the same as shown lite is influenced significantly by the pretreatments the in Table 2. The capacity of the conditioned zeolite, zeolite has received and the influence of various pre1.69 + 0.08 meq/g dry zeolite, was smaller than those treatments is poorly documented. In addition, zeolites other than clinoptiiolite have not been evaluated for of the reconditioned and acid-washed zeolite, 1.82 + 0.23 and 1.86 + 0.05 meodg dry zeolite, re- removing ammonium from wastewater by ion spectively. Although the latter two preparations gave exchange. Hayhurst (1976) evaluated mordenite, philapproximately the same average capacity, the varia- lipsite and clinoptilolite for removing NH3 from gas bility in the value of the reconditioned zeolite was streams. Phillipsite was demonstrated to be best in much greater than that of the acid-washed zeolite. removing NH3 from gas. This paper presents the results of some laboratory Reconditioning with a salt solution adjusted to pH 10.0 should have removed acidic impurities from the studies to evaluate phillipsite, mordenite, erionite and zeolite. Similarly, acid-washing should have removed pretreated clinoptilolites for ammonium removal from wastewater. Ammonium capacities were any alkali impurities. When capacity tests on the same zeolite sample measured in two ways: (1) using a 100mg/l solution were repeated following regeneration with 1 N NaCI, of NH,~--N as a feed solution in column exhaustion the zeolite capacities were observed to increase. After tests and (2) using a synthetic sewage containing repeating the test three times the capacity appeared to 14mg/l N H 2 - - N as a feed solution in column exhaustion tests. Packed bed densities were deterlevel off at 2.24 meq/g dry zeolite. Jorgensen et al. (1976) also noted that the capacity mined to convert observed capacities on a weight of of a sample of clinoptilolite increased with the zeolite basis to capacities on a volume of zeolite basis. number of regenerations. Their regeneration process A crushing test was used to assess physical differences consisted of treatment with 4% NaOH for 30 min, in the zeolites. followed by a rinse with 11 of deionized water for 40 min. After repeating this procedure four times, the measured capacity was found to be constant. They LABORATORY TECHNIQUES AND EQUIPMENT concluded that the sodium hydroxide "activates" the clinoptilolite. The results of Seyforth's study suggest Grading and pretreating the zeolites Twenty-five pound samples of erionite, mordenite, that it was not necessarily the sodium hydroxide which was responsible for the increased capacity. It and phillipsite were provided by Anaconda Company, may, however, be related to the number of exchange Denver, Colorado for use in this study. Clinoptilolite cycles that the zeolite is exposed to after being oven from Buckhorn, New Mexico was provided by dried at 103°C. This may be seen by comparing the Double Eagle Mining Co. (Casper, Wyoming) and
Pretreated natural zeolites
"F-, .~ ~
~g
2~
0
"" ,~
163
z'~z~ o ~
.~:-~
~:'~'.="-~,,~.o~oo~. •- + 0
.~.~
~;
.~'~o-0
~o
= ~
~
o
0 0
e-
[-
.~__ .o .o ._.
~. ~.=_
_q
Z
0 'n ~
0 '."
Z
0 0 ,m
l-
g
~
~,~ o~
Z
X
X
X
X
o
o
~
o
8
~o +~
Z
o
~8
n-
8
e~
o
o ÷
F..
o
..0 n~
,.o'~
"~ ~
+~
+~
.+_+~
Zo~
Z
Z
o
0 0
g.
0 ..~,
0
e~
0
g
g u~
o ..d 0
0
164
J.R. KLH!Vl and M. J. SEMMENS Table 2. Results of the capacity tests conducted by Seyforth (16) Type of clinoptilolite sample
Number of samples tested
Capacity meq g t
Sample test number
5 3 2 6 2 4*
1.69 + 0.08 2.11 +_ 0.10 2.23 + 0.01 1.82 + 0.23 2.25 + 0.07 1.85 + 0.05
1 2 3 I 2 1
Conditioned Conditioned Conditioned Reconditioned Reconditioned Acid washed
* Capacities measured by both ammonium and sodium elution studies. graded with U.S. standard mesh sieves. The 20 × 30 mesh fractions were retained for use in this study. Zeolite samples were pretreated in different ways. Solutions were pumped from a feed reservoir upflow through a 2.5 cm acrylic column containing approximately 80 g of a zeolite sample. A nylon screen supported the zeolite in the column. Table 3 presents the various pretreatment conditions employed to prepare the zeolite samples. All samples were treated with approximately 41 of l N NaC1 over a 2 h period. Some samples were given two treatments. Following NaCI treatment zeolite samples were washed with deionized water. Nitric acid treatment was accomplished by contacting a zeolite sample with 21 of 1 N HNO3 over a l h period. Samples 5 and 5B were contacted with l N N a O H treatments of 21 in 70 min and 31 in 1 l0 rain, respectively. The heat treated zeolite samples were heated to 60ffC for 1 h, cooled overnight and rewetted. Some samples were autoclaved at 30psi g 134°C for 1 h.
tion of ammonium chloride (100mg/l NH~--N) in deionized water. The feed solution was supplied to the column from an elevated reservoir at a flow rate of 20 bed vois/h (10 ml/min), 1 1 samples of the column effluent were collected and analyzed for ammonium concentration. Samples of the feed were analyzed for ammonium concentration each time an effluent sample was collected to ensure there was no change in the feed ammonium concentration during the test period. When the effluent ammonium concentration was the same as that of the feed (within analytical error), the zeolite was flushed with deionized water, removed from the column, dried and weighed. Total operating capacity measurements. These measurements were made using exactly the same procedure employed for total capacity measurements except that the feed solution had the composition presented in Table 4. This solution composition was identical to that used by Breck (1973) and contained no organics or colloids.
Testing the zeolite,s Measurements were made of the total ammonium capacities of the pretreated zeolites. In addition, their capacities for ammonium in the presence of competing ions were measured to determine any improvement in selectivity. Total capacity measurements. A 50 ml burette containing 30ml of pretreated zeolite supported on a glass wool plug was contacted downflow with a solu-
Analytical techniques Ammonium measurements were made using an Orion Model 95-10 ammonia selective electrode connected to an Orion Model 701 digital pH/mV meter (Orion, Cambridge, Mass.). The electrode tended to drift and standard ammonia samples were analyzed between every analysis of an unknown sample as a check on the calibration curve. Duplicates of the an-
Table 3. Pretreatment scheme for clinoptilolite samples tested
Sample Clinoptilolite I Clinoptilolite I B Clinoptilolite 2 Clinoptilolite 2B Clinoptilolite 3 Clinoptilolite 3B Clinoptilolite 4 Clinoptilolite 5 Clinoptitolite 5B Phillipsite Mordenite Irionite
HNO3 + rinse
NaOH + rinse
Second NaCI + rinse
First NaCI
Backwash
X
X
X
X
X
×
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
×
X
X
X
X
X
X
Heat treatment Autoclaving
X X X
165
Pretreated natural zeolites Table 4. Composition of the synthetic waste water (14) Element
Concentration
Compound used
NH2 Na + K+ Mg + + Ca + + pH
14 mg/l 58 mg/l 12 mg/l
(NH4): SO4 NaCI, NaHCO3 KC1 MgSO4 CaCI2.2H20 Adjusted with proper ratio of NaHCO3 to NaCI
8 mg/I
34 mg/l 7.45
Table 5. Capacity of zeolites using 100 nag NH~'--N/I solution as the feed solution Capacity (meq NH+--N/g) Individual tests Mean
Zeolite Clinoptilolite 1 Clinoptilolite 2 Clinoptilolite 3 Clinoptilolite 5 Phillipsite Mordenite Erionite
1.70, 1.51, 1.62, 1.85, 2.73, 1.76, 0.30
known samples and occasionally triplicates of the unknown samples were run when the drift was excessive. To weigh the zeolites in a reproducible way wet zeolite samples were dried in a 100°C oven overnight and then placed in a desiccator over a saturated solution of NaC1. The desiccator was placed in a constant temperature (37°C) oven to maintain a constant humidity (75%) environment (15). The samples were held in the desiccator for 2-3 days to assure equilibrium with the humidity in the desiccator.
Zeolite packed bed densities Since the specific gravity and packing characteristics of the zeolites vary, packed bed densities were measured in order that volumetric capacities could be determined. Volumetric capacities are important for sizing ion exchange equipment. A 25 ml graduated cylinder filled partially with water was filled to 25 ml with wet zeolite. The cylinder was rapped on the lab counter until the zeolite volume showed no further reduction~ Final volume was recorded. The samples were removed from the cylinder, dried and weighed.
Zeolite strenyth To compare the physical strength of the zeolite samples a crushing test developed by Hayhurst (1976) was used. Zeolites employed in column operations are prone to physical attrition during backwashing and regeneration cycles. It is desirable, therefore, to compare the physical strength of the zeolites in order to determine their suitability for column operation. A single zeolite grain was placed on the weighing pan of a top loading direct reading Mettler Balance (Princeton, New Jersey). The balance was positioned beneath a drill press fitted with a 1/4 in steel rod. The
1.61, 1.61, 1.50, 1.73, 2.55 1.83
1.60, 1.76, 1:57, 1.75,
1.61, 1.64 1.67, 1.65 1.57 1.75
1.63 1.64 1.57 1.77 2.64 1.80 0.30
rod was slowly lowered on to zeolite particle and pressure was applied gradually until the grain fractured. The weight required to crush the particle was recorded as a measure of crushing strength. The test was run on 20 grains of each zeolite sample to obtain a significant mean value for comparative purposes.
X-ray diffraction analyses Standard X-ray diffraction patterns were obtained for powdered samples of untreated and pretreated zeolites using a Phillips-Norelco Diffractometer. The instt'ument employed CuK ~ radiation at 1° 20/rain. RESULTS The results of the capacity tests are presented in Tables 5 and 6. The test techniques ensured excellent reproducibility as indicated by the standard deviation value§. Clinoptilolite samples 1, 2 and 4 all had an average total ammonium capacity of 1.63-1.64 meq/g. Sample 3 treated clinoptilolite had a somewhat lower capacity of 1.57 meq/g and sample 5, the sample that received a treatment of sodium hydroxide, had the highest observed capacity of 1.77 meq/g. Table 6. Capacity of zeolites using synthetic sewage as the feed solution Zeolite
Capacity (meq NH2--N/g) Individual tests Mean
Clinoptilolite 1 Clinoptilolite 2 Clinoptilolite 3 Clinoptilolite 5 Phillipsite Mordenite
0.499, 0:502, 0.498 0.500 0.530, 0.514, 0.526 "0.523 0.590, 0.594 0.592 0:500, 0.514 0.507 0.970, 1.031 1.001 0.439, 0.444 0.442
166
J.R. KLIEVI~and M. J. S~MM~:NS
Table 7. Packed bed densities of 20 × 30 mesh zeolite in Na ÷ form
batch of clinoptilolite. The reason for this difference in measured capacities may stem from the different techniques employed for measuring the weight of zeoPacked bed density lite. Seyforth dried the zeolite at 105°C, cooled it in a Zeolite g dry zeolite/submerged ml desiccator and weighed the dry zeolite. In this study it Clinoptilolite 1 0.848 was observed, however, that this technique could Clinoptilolite 2 0.844 cause significant variation in the measured weight of a Clinoptilolite 3 0.853 zeolite sample. Natural zeolites are extremely effective Clinoptilolite 4 0.849 dessicants themselves and the dried zeolite samples Clinoptilolite 5 0.857 Mordenite 0.699 compete with chemical dessicants for the available Phillipsite 0.534 water. Thus, if the dessicant used is partially Erionite 0.503 exhausted, the zeolite may extract some of the available water and as a result weigh heavier than when Mordenite had a similar capacity to clinoptilolite fresh dessicant is used. The preferred equilibration but erionite had a very much lower total exchange with a constant, high humidity atmosphere gives capacity of only 0.3 meq/g. Phillipsite, with a capacity more reproducible results but the absorbed water tends to make the weight capacity appear smaller. of 2.64 meq/g was clearly superior to clinoptilolite. Tests on the ammonium capacities measured folAs noted above, differences between the clinoptilolowing exposure to synthetic sewage indicated that lite sample studied and other zeolites preclude direct the heat treated clinoptilolite (sample 3) was most comparison of these results with those reported by selective for ammonium. The other clinoptilolite other investigators, but the values obtained (Table 5~ samples showed that the other pretreatment tech- do appear similar to those reported by others (Table 1). niques had little impact on ammonium selectivity. In this study phillipsite was observed to have a Phillipsite was equally selective for ammonium and weight capacity that was 70°,0 greater than that of its higher total capacity resulted in an effective capa- similarly treated clinoptilolite. Tests conducted on the city that was about double that of clinoptilolite on a selective removal of ammonium from a synthetic sewweight basis. age containing competing ions indicated that the philThe packed bed densities of the zeolite samples are lipsite was more selective for ammonium than clinoppresented in Table 7. The type of pretreatment had no tilolite and removed twice as much as ammonium on effect on the packed density of clinoptilolite which a dry weight basis. However, the low packed bed denremained at a value of approximately 0.85g dry t sity of the phillipsite compared to that of clinoptiiolite zeolite/ml. The phillipsite was considerably lighter, reduced the performance of phillipsite on a voluhaving a packed density of only 0.53 g dry zeolite/ml. metric basis. Its volumetric capacity was only 26~, The crushing tests indicated that samples I and 3 of better than similarly treated clinoptilolite for clinoptilolite had the highest measured crushing ammonium removal from synthetic sewage. Phillipsite has two disadvantages that must instrengths. The results are depicted in Table 8. The other pretreatments for clinoptilolite reduced the fluence its use in water treatment, however. Firstly, strength of the zeolite. Phillipsite was observed to be this zeolite is structurally very weak and breaks easily. extremely friable and yielded a very low crushing Even in limited laboratory studies physical attristrength. The standard deviations were poor for this tion of the phillipsite was significant and the fines test and the difference between samples 1 and 3 of produced created large headlosses in the column clinoptilolite reported in Table 8 cannot be con- studies. Secondly, phillipsite is considerably more expensive than clinoptiiolite as shown in Table 9. sidered significant. Possibly the attrition problem could be overcome by mixing powdered phillipsite with a porous binder DISCUSSION OF RESULTS but as can be seen the costs of the zeolite would not The observed capacities of the pretreated clinopti- be competitive with clinoptilolite. However, if lolite samples tested in this study appear a little lower through the use of a binder the packed bed density of than those reported by Seyforth (1975) for the same the zeolite could be increased without reducing the
Table 8. Results of crushing test Zeolite Clinoptiiolite 1 Clinoptilolite 3 Phillipsite Mordenite Erionite
Number of tests
Mean crushing strength (g)
Standard deviation
20 20 20 20 20
366 329 83 129 105
114 lit 63 100 51
G
Pretreated natural zeolites Table 9. Approximate cost of zeolites in bulk quantities (22) (FOB) Zeolite Clinoptilolite Phillipsite Mordenite Erionite
S/ton 200-210 40(0450 300 400
exchange capacity significantly, then this may prove feasible. Analysis of the volumetric capacities of the clinoptilolite indicated that the heat treated sample was the best for ammonium removal from synthetic sewage. The heat treated clinoptilolite had approximately 17% more capacity than clinoptilolite that had been exposed only to NaCI. This pretreatment improved the selectivity of the zeolite for NH2 in exchange with competing cations and its measured capacity approached that of phillipsite. The heat treatment caused the clinoptilolite to turn dark brown, suggesting that iron oxidation had occurred within the zeolite structure. Why this should improve the zeolite's selectivity for NH,~ is not clear. Studies conducted by Anaconda have shown that the iron content of clinoptilolite is not exchangeable nor is it leached by applied EDTA (personal communication, 1978). The X-ray diffraction studies on the variously treated clinoptilolite samples unfortunately provided no evidence of any structural changes. Indeed, all the X-ray diffraction scans on the pretreated samples were identical. Further studies are needed to characterize the influence of heat on the zeolite. In this study a fixed time at a fixed temperature was examined. The time and temperature of pretreatment may be important. In addition, the technique of rehydration may be influential in affecting zeolite performance. During the course of pretreatment the actual weight of a zeolite sample may change. Thus, capacities reported per unit weight of the pretreated zeolite may be misleading. For example, if pretreatment reduces the weight of a zeolite sample without influencing capacity, then naturally the exchange capacity/unit weight will appear higher. Yet no great advantage may be realized from such a pretreatment. In this study weight changes were recorded for different pretreatments. No significant weight changes were recorded apart from the following: 1. Heat treatment caused a 4% weight loss. Since the total exchange capacity also declined by 4% following heat treatment, it would appear that some exchange sites are lost during heat pretreatment. 2. NaOH pretreatment caused a slight weight gain, 2%. The improved total exchange capacity of the base pretreated zeolite must therefore stem from the exposure of more exchange sites. 3. Acid pretreatment caused a 9% weight loss and no impact on the exchange capacity which suggests
167
that zeolite structure itself may be broken down by the strongly acidic conditions employed in pretreatment. The ion exchange process is a diffusion controlled process and the kinetics of exchange may be influential in the choice of a zeolite where capacity and selectivity are more or less equal. Some preliminary data was collected from the column studies on the pretreated clinoptilolite samples that suggested that the kinetics of exchange were not significantly influenced by pretreatment. Additionally, the kinetics of exchange with phillipsite were comparable with clinoptilolite.
CONCLUSIONS several pretreated clinoptilolites and other natural zeolites have been evaluated for ammonium removal from wastewater. The following conclusions may be drawn from this study: 1. The weight of a natural zeolite is influenced significantly by the relative humidity. It was therefore necessary to weight zeolites equilibrated with a constant relative humidity in order to obtain accurate and reproducible capacity data. 2. By heat treating clinoptilolite, the zeolite's capacity for NH2 was not significantly affected but the zeolite capacity for N H 2 - - N in the presence of competing cations was observed to increase. The selectivity of the zeolite for ammonium was thus enhanced. 3. By treating the clinoptilolite with NaOH, capacity for N H 2 - - N using a pure solution of NH4CI as feed was slightly improved, but the zeolite's selective capacity for N H 2 - - N in the presence of competing cations was not affected. 4. Acid and steam treatments did not influence clinoptilolite's capacity for N H 2 - - N . 5. Erionite, mordenite and phillipsite samples (Anaconda) were also tested although no pretreatments were applied to these other than conversion to the sodium form by 1 N NaCI. The phiihpsite sample was very effective in ammonium removal but erionite and mordenite'behaved poorly. The phillipsite proved to be better than even pretreated clinoptilolite samples in both weight and volumetric capacities. Unfortunately, phillipsite.is structurally weak and breaks down easily to produce fines. This will limit application of this zeolite in water treatment. If the strength of the zeolite could be improved through the use of a binder, it may find a wider application.
REFERENCES
Ames L. L. (1962) Effect of base cation on the cesium kinetics of clinoptilolite. Am. Min. 47, 1310-1316. Ames L. L. (1964) Some zeolite equilibria with alkali metal cations. Am. Min. 49, 127-145.
168
J.R. KLniVl! and M. J. SEMMENS
Ames L. L. (1967) Zeolitic removal of ammonium ions from agricultural and other wastewaters. 13th Pacific Northwest Industrial Waste Conference Proceedings, Washington State University, Pullman, Washington. Barrer R. M., Papadopoulos R. & Rees L. V. C. (1967) Exchange of sodium in clinoptilolite by organic cations. J. lnor 9. Nucl. Chem. 29, 2047-2065. Breck D. W. (1973) U.S. Patent 3,723,308. Dean J. A. (19731 Lange's Handbook qf Chemistry. l lth Edition, McGraw-Hill, p. 10-79. Deffeyes (1959) Anl. Min. 44, 501. Frysinger G. R. (1962) Cesium-sodium ion exchange on clinoptilolite. Nature 194, 351 353. Harris & Brindly (1954) Am. Min. 39, 819-824. Hayhurst D. T. (1976) Potential use of natural zeolites for ammonia removal during coal gasification, presented at Zeolite 1967 Cot!ference, Tucson, Arizona. Hoss & Roy (1960) Beitr. Minerl. Petro,q. 7, 389. Howery D. G. & Thomas H. C. (1965) Ion Exchange on the mineral clinoptilolite. J. Phys. Chem. 69(2), 531 537. Jorgensen S. E. (1976) Ammonia removal by use of clinoptilolite. Water Res. 10, 213-224. Mumpton F. A. (1960) Clinoptilolite redefined. Am. Min. 45, 351-369.
Murphy C. B., Hrycyk O. & Gleason W. T. (19761 Natural zeolites: novel uses and regeneration in wastewater treatment, presented at Zeolite 1976 ConiC,fence, Tucson, Arizona. Obretenov Z., Dimitrove D., Todorova V. & Toncheva E. 11976) Production of absorbents with molecular sieve action from native bulgarian clinoptilolite. Wiss. Z. Tech. Hachsch. "'Carl Schorlemmer'~ Lcuna-Mersehur¢4, 18{4), 569-572. Personal communication with Mr. Dennis Leonard, Anaconda Company, 22 August 1977. Personal communication with Mr. Dennis Leonard, Anaconda Company, December 2, 1977. Personal communication with Dr. Harold Vincent, Anaconda Company, 2 December 1977. Personal communication with Dr. Harold Vincent, Anaconda Company, 30 August 1978. Seyforth M. C. (19751 The Selectivity c~f Clinoptilolite Jor Heary Metal Cations, M.S. Thesis, University of lllinois, Urbana. Weast R. C. (1974)(Ed.) Handbook oJ Chenlistry and Physics, 55th Edition. The Chemical Rubber Company Press, p. 10-79.