Chemically activated carbon on a fiberglass substrate for removal of trace atrazine from water

Abstract

Chemically activated fiber (CAF) for removal of trace atrazine from water was prepared by coating fiberglass assemblies with a phenolic resin along with a chemical activation agent of ZnCl₂, then stabilization and heat treatment in N₂ at 500 °C. The carbon on the CAF shows similar BET surface area and volumes of narrow micropores (<10 Å), higher volumes of wide micropores (10–20 Å) and narrow mesopores (20–50 Å), as compared with a commercially available GAC F-400. Adsorption isotherm data show that the CAF has a higher adsorption capacity for atrazine than the GAC, primarily because the CAF has an increased pore (10–50 Å) volume. The breakthrough tests show that the CAF filter is ten times more effective over the GAC filter in removing the atrazine to below the current USEPA standard of 3 ppb. The CAF filter also shows a better competitive adsorption of atrazine over the GAC filter in the presence of 50 times higher concentration of humic acid. Such a filter can be regenerated to 90% of its original activity by heating at 350 °C in air.

---

Introduction

It is well known that atrazine is the most widely used herbicide in the United States, with more than 80 million lbs sold annually.^1–3 Past studies have built a strong case that atrazine is unsafe for humans. Some of these studies have found that the herbicide disrupts the production of normal human hormones, while others have concluded that it is very likely associated with a higher incidence of cancer in both humans and laboratory animals.^4–8 Because of the fact that atrazine is found in drinking water across the country, the USEPA has set a maximum contaminant level (MCL) for atrazine at 3 parts per billion (ppb).

The use of activated carbons is the treatment method designated by USEPA as the best available technology for removing atrazine from drinking water.⁹ Numerous studies have reported on the adsorption of atrazine using conventional activated carbon granules and fibers.^10–13 It has been found that the pore size and pore size distributions of activated carbons have a strong effect on the adsorption of atrazine in water.^11–13 The presence of natural organic matter (NOM) can have a deleterious effect on the ability of activated carbon to remove trace atrazine, in terms of both adsorption kinetics¹² and adsorption capacity,^11,14,15 due to pore blocking by NOM and direct adsorption competition. It was proposed that mesopores could alleviate pore blockage¹⁶ and maintain a fast kinetic adsorption rate.¹²

A new generation of adsorption fibers with low cost, and high mesopore content, has been developed in our laboratory by coating a glass fiber with polymers in a solvent along with a catalyst and then activating at relatively low temperatures.¹⁷ This method which is referred to as chemical activation provides the fibers with much higher yields, high surface areas, higher mesopore volumes and some unusual pore surface chemistries.^18–21 The previous studies have shown that activated carbon on fiberglass has a higher removal efficiency than granular activated carbon (GAC).^22,23

In the present work, chemically activated fiber (CAF) was prepared to remove trace atrazine from water. The differences in pore structures and surface chemistries between the CAF and the GAC were investigated. The adsorption isotherms and the breakthrough tests on the CAF and GAC were carried out to compare their respective adsorption abilites for trace atrazine in water. The thermal regeneration of the CAF filter was evaluated to determine the potential for reuse of such a filter.

Experimental

Materials

Atrazine PESTANAL^® (2-chloro-4-ethylamino-6-isopropylamino-1,3,5-triazine) with assay of 99.2% was obtained from Sigma-Aldrich. Humic acid (sodium salt) with a melting point >300 °C was obtained from Aldrich Chemical Company. Phenolic resin GP 2006, obtained from Georgia Pacific, was employed as the starting material for the preparation of the adsorbent used in this study. Zinc chloride (98+%) and ethyl alcohol (denatured), obtained from Aldrich, were used as a chemical reagent and a solvent, respectively. Hydrochloric acid (37.2%) and deionized (D. I.) water were used to wash the samples. The substrate fiber was a non-woven fiberglass mat, Craneglas^® 230 (0.015 nominal, fiber diameter of 6.5 µm, obtained from CRANE & CO. (Dalton, MA)). As a point of comparison, granular activated carbon GAC Filtrasorb 400 (effective size: 0.55–0.75 mm) was employed. This GAC is made by Calgon Carbon Corporation for the removal of taste and odor compounds and dissolved organic compounds in potable water treatment. It can be used to treat surface and ground water sources.

Preparation of filter assembly cartridges

Phenolic resin (45 g) and ZnCl₂ (45 g) were dissolved in 300 ml of ethanol to give a fluid mixture. The fiber filter assembly cartridge was prepared as shown in Fig. 1. A glass mat was dip-coated with the mixture and dried in air at room temperature. A roll of impregnated glass mat was formed, and then dried and cured in air at 100–170 °C for 6 h. The cured roll was transferred to a separate furnace purged with flowing N₂ and activated at 500 °C for 30 min. After cooling in flowing N₂, the chemically activated fiber (CAF) filter cartridge was thoroughly washed with D. I. water, 0.5 M HCl and then rinsed with D. I. water. After that the cartridge was first dried in an air convection oven at 150 °C for 1 h, and then transferred to a vacuum oven and dried further at 150 °C under vacuum for at least 12 h.

Fig. 1 The preparation of a CAF filter assembly cartridge.

The CAF filter was then assembled as part of the above cartridge, which included glass tubing with two ends sealed, two plastic connectors and parafilm wrapped around the CAF filter cartridge. The schematic drawing of the CAF filter is shown in Fig. 2. In comparison, a GAC filter containing an equal weight of adsorbent, shown in Fig. 2 (GAC filter-1), was constructed from plastic tubing with a thin glass mat at the bottom to hold the GAC in place and loose glass fibers at the top to take up free space. Another GAC filter-2 containing an equal volume of adsorbent was also constructed from plastic tubing .The physical characteristics of the filters for breakthrough profile experiments are listed in Table 1.

Fig. 2 Schematic drawings of CAF and GAC filters.

Table 1 Physical characteristics of filters for breakthrough profile experiments on CAF filters and GAC filters

	Equal weight		Equal volume
	CAF filter-1	GAC filter-1	CAF filter-2	GAC filter-2
Carbon coating (wt.%)	50	—	48	—
Adsorbent size/µm	6.5 + coating	550–750	6.5 + coating	550–750
Bed mass/g	4.2	4.2	4.0	4.5
Length of bed/mm	135	77	136	136
Diameter of bed/mm	13.9	15.7	13.5	10
Core/mm	8.95	8.95	8.95	None
Cross sectional area/mm^2	88.8	130.7	80.2	78.5
Volume/cm^3	12	10	11	11
Bulk density/g cm^−3	0.35	0.42	0.36	0.41
Flow rate/bed volume h^−1	200	240	218	218

---

The thermal regeneration of the CAF filter was carried out by heating the CAF filter cartridge at 350 °C in air after peeling off the parafilm and taking off the two plastic connectors.

Adsorption of atrazine

Adsorption isotherms. A high concentration atrazine solution (about 19 mg L⁻¹) was first made by dissolving 19 mg of atrazine into 1 L of D. I. water. This solution could quickly be diluted to desired concentrations. CAF samples used for isotherm experiments were fabricated in the same manner as discussed previously, except that fiber assemblies were not rolled into filter cartridge form. Static adsorption experiments were carried out by placing equal weights (∼20 mg of CAF or ∼10 mg of GAC) of samples into each of 40 ml vials sealed with a septum cap and then adding aqueous solutions of atrazine of varied initial concentration (500–19000 ppb). After the vials were fully prepared, they were placed in a Millipore Rotary Agitator and tumbled for 1 week at 25 °C. The initial and residual concentrations for each sample were measured after 1 week. The atrazine adsorbed (mg g⁻¹) was calculated from the difference in concentration between the initial (C₀) and residual or equilibrium (C_e) solutions.

Breakthrough tests. All the filters were put into water and the air bubbles inside the filter were driven out under vacuum prior to use.

The experimental apparatus for breakthrough tests contains a stock (12 L) with pre-prepared atrazine solution (∼100 ppb) and a cartridge pump. For the competitive experiments, 60 mg of humic acid (sodium salt) was added into the stock to prepare a mixed solution with 5000 ppb of humic acid and 100 ppb of atrazine. Both the CAF and GAC filters containing equal weights or equal volumes of adsorbent were evaluated at the same time. A flow rate of 40 ml min⁻¹ of contaminant solution was controlled with a cartridge pump and passed through the two filters. The solution was sampled every 30 min for effluent and 60 min for influent.

Atrazine analysis. The concentration of atrazine was analyzed by a Hewlett Packard 5890 Gas Chromatograph (GC) equipped with an Electron Capture Detector (ECD). For lower concentrations, a Zymark TurboVap^® 500 was used to pre-concentrate the atrazine prior to analysis using a GC-ECD.

Characterization

Gas adsorption. The adsorption of nitrogen at 77 K was carried out with an Autosorb-1 volumetric sorption analyzer controlled by Autosorb-1 software (Quantachrome Corp.). All samples were degassed at 150 °C until the outgas pressure rise was below 5 μmHg min⁻¹ prior to analysis. N₂ isotherm results in the appropriate relative pressure ranges were used for subsequent calculations. The BET surface areas were determined using 5 points (P/P₀ = 0.01, 0.015, 0.02, 0.025, 0.03) from the N₂ adsorption isotherm at 77 K. The total pore volume was estimated from the amount of nitrogen adsorbed at P/P₀ = 0.95. A method based on the density functional theory (DFT) NLDFT model, provided by the Autosorb-1 for windows 1.50 software (Quantachrome Corp.), was applied to the N₂ adsorption data at 77 K to get pore volumes in the pore size ranges of <10 Å, 10–20 Å and 20–50 Å. The volume of mesopores (20–500 Å) was calculated by subtracting the volume of micropores (<20 Å) from the total pore volume at a relative pressure of 0.95.

Elemental analysis (EA). A Model CE440 Elemental Analyzer (EA) was used to directly determine the C, H, and N weight percentages in the samples. The oxygen content was calculated by mass difference after combining the results of TGA and assuming that only C, H, N, O, and glass for CAF and C, H, N, O, and ash for GAC were present and the glass weight remained unchanged after TGA burn-off in air at 700 °C.

Thermogravimetric analysis (TGA). The amount of the coating deposited on the fiberglass mat was measured using a Hi-Res TA Instruments 2950 Thermogravimetric Analyzer by burning off the coating in air at 700 °C. TGA analysis of atrazine was done in air or N₂ using a heating rate of 10 °C min⁻¹ to 300 °C.

Results and discussion

Characterization of CAF and GAC

ZnCl₂ activated phenolic resin on a fiberglass substrate (or chemically activated fiber CAF) was compared with a commercially available GAC F-400 in this study to elucidate the adsorption effectiveness of atrazine on the CAF. The characteristics of the porous structures of CAF and GAC are listed in Table 2. The results show that the carbon on the CAF has similar BET surface area and volume of narrow micropores (<10 Å) as the GAC. However, the carbon on the CAF has higher volumes of wide micropores (10–20 Å), narrow mesopores (20–50 Å), total mesopores (20–500 Å), and total pores than the GAC. In general, this chemically activated carbon showed higher pore volumes over the GAC F-400 in the range of pore sizes between 10 and 50 Å which might be a suitable size for the adsorption of atrazine because the molecular size of atrazine is around 3 × 8.4 × 9.6 Å.²⁴

Table 2 Porous structure characteristics of activated carbons (based on carbon)

Sample	BET surface area/m^2 g^−1	DFT method pore volume/cm^3 g^−1			Mesopore (20–500 Å) volume/cm^3 g^−1	Total (P/P0 = 0.95) pore volume/cm^3 g^−1
		<10 Å	10–20 Å	20–50 Å
CAF	1098	0.21	0.16	0.14	0.24	0.61
GAC-F400	1009	0.22	0.13	0.11	0.19	0.54

---

The elemental analysis results of CAF and GAC are shown in Table 3. In order to obtain accurate values the carbon weight percent was measured using thermogravimetric analysis (TGA) by burning off the coating in air. The numbers are displayed for both the GAC and the CAF normalized to the carbon coating to allow for easier comparison. The carbon on the CAF shows 28 times higher atomic ratio of hydrogen/carbon (H/C) and slightly higher atomic ratio of oxygen/carbon (O/C) than the GAC. This result indicates that the coating on the CAF has a low level of carbonization due to the low temperature treatment at 500 °C. Thus, the carbon on the CAF has a different material chemistry from the GAC.

Table 3 Elemental analysis of activated carbons

Sample	C (wt%)	H (wt%)	N (wt%)	O^a (wt%)	H/C atomic ratio	O^a/C atomic ratio
a Data calculated from the results of elemental analysis (EA) and TGA. The elemental analyses of C, H and N have associated errors of ±0.40%, thus the % O results determined by difference are subject to a cumulative error. b TGA showed an ash content of 5.97 wt%.
CAF	89.55	2.58	0.43	7.45	0.346	0.062
GAC-F400^b	88.27	0.09	0.69	4.97	0.012	0.042

---

TGA analysis in Fig. 3 shows that the CAF has around 50 wt% of carbon coating and is thermally stable in air until 400 °C, suggesting that a high temperature up to 400 °C can be used to treat such a CAF in air.

Fig. 3 TGA analysis of the CAF in air.

Adsorption isotherms

The Freundlich isotherm is a very common expression employed to describe the adsorption of atrazine onto activated carbon.^11,24 The univariate expression, modeling adsorption of one contaminant from water, has the form of a power model:

Q = KC_e^1/n (1)

where Q = mass of the target compound adsorbed per unit mass of adsorbent (mg g⁻¹); C_e = equilibrium concentration of the adsorbable compound in the liquid (µg L⁻¹); K = adsorption equilibrium constant (mg g⁻¹)(µg L⁻¹)^−1/n; and n is a constant indicative of adsorption intensity.

Based on this equation, adsorption isotherms of atrazine onto CAF and GAC were obtained by plotting the data on log₁₀ graph. The scattered data are shown in Fig. 4 along with the best-fit lines, calculated adsorption parameters K, 1/n values, and the correlation coefficient values (r). For an effective adsorption system, high adsorption capacity (Q) is needed. Thus, if the C_e is not lower than 1 ppb, the larger the K and 1/n values, the more effective the carbon is for adsorption.

Fig. 4 Atrazine adsorption isotherms on CAF and GAC.

Adsorption isotherms and parameters based on the Freundlich equation (Fig. 4) show that the CAF has higher atrazine adsorption capacities than the GAC. This can be explained by 1) CAF's higher pore volumes in the pore size range between 10 and 50 Å whereby the atrazine molecule is easy to enter such a porous structure; and 2) the higher H/C ratio in CAF which might improve the adsorption of atrazine from water.

Breakthrough experiments

The breakthrough tests (see Fig. 5) on equal weights of CAF and GAC filters (see Table 1) running at a flow speed of 40 ml min⁻¹ show that the CAF filter-1 is ten times better, in a comparable number of bed volume of the GAC filter-1, in removing trace atrazine (∼100 ppb) to below the current USEPA standard (3 parts per billion). It is even more effective at breakthrough levels below 1 ppb. This is not unexpected since the fiber (0.65 µm + coating) form of the CAF offers excellent contact efficiency along with fast adsorption kinetics compared to the large particle size (0.55–0.75 mm) and same weight of the GAC.

Fig. 5 Breakthrough tests for atrazine comparing CAF filter-1 with GAC filter-1.

To further compare the CAF filter with the GAC filter in removing trace atrazine from water, the breakthrough tests using equal bed volumes of CAF filter-2 and GAC filter-2 (see Table 1) were carried out and the results are shown in Fig. 6. In this experiment, the influent solution contains ∼100 ppb of atrazine and ∼5000 ppb of humic acid. The CAF filter-2 has less mass of adsorbent than the GAC filter-2 which contains more GAC than GAC filter-1 (see Table 1). The experimental results show the GAC filter-2 cannot remove trace atrazine down to 3 ppb in the presence of a high concentration of humic acid, indicating that the presence of humic acid greatly decreases the adsorption kinetics of trace atrazine onto the GAC. It was reported that natural organic matter adversely affects both the equilibrium and kinetics of trace organic compound adsorption by activated carbon through two major competitive mechanisms: direct site competition and pore constriction/blockage.¹⁶ However, the CAF filter-2 displays a much higher breakthrough number of bed volumes for the removal of trace atrazine from water in the presence of a 50 times high concentration of humic acid. These values indicate that the CAF filter has much faster adsorption kinetics in selectively removing trace atrazine from water. They also suggest that the following features may be advantageous: 1) the fiber form displays a higher concentration of available pores; and 2) there is a higher content of pore volume in the pore size range between 10 and 50 Å.

Fig. 6 Breakthrough tests for atrazine comparing CAF filter-2 with GAC filter-2 in the presence of a high concentration of humic acid.

Regeneration of CAF filter

According to the description of atrazine properties from Sigma-Aldrich, atrazine has a melting range from 176.7 to 177.8 °C. TGA analysis of atrazine in N₂ or in air (see Fig. 7) also shows that atrazine disappears when the temperature increases above 200 °C, due to a sublimation/evaporation process and /or possible destruction of atrazine. This suggests that the CAF can be thermally regenerated. In a previous study it was shown that PVA and PAN-based, chemically activated fibers containing adsorbed atrazine could be fully regenerated by heating the filter at 190 °C under vacuum.²³

Fig. 7 TGA analysis of atrazine in N₂ or air.

In the case of competitive adsorption of atrazine and humic acid on the CAF filter, regeneration is very difficult since the humic acid used in this study has a melting point >300 °C. It was found that the BET surface area of the CAF decreased from 1098 to 832 m² g⁻¹ of coating after the adsorption of humic acid, but it recovered to 1088 m² g⁻¹ after heat treatment in air at 350 °C for 30 min. TGA also shows that the CAF has high stability up to 400 °C in air (Fig. 3). Thus, a temperature of 350 °C in air was chosen to regenerate the CAF filter-2 in this study. The breakthrough tests (Fig. 6) on the regenerated CAF filter-2 show that the CAF filter-2 can be regenerated to 90% of its original activity. It is predicted that the thermal regeneration efficiency could be higher than 90% if the influent atrazine solution contains a low concentration of humic acid.

Conclusions

Chemically activated fiber (CAF) was prepared from phenolic resin on a fiberglass substrate with ZnCl₂ activation at a temperature of 500 °C. The CAF was compared with a commercially available GAC F-400 for removal of trace atrazine from water. The carbons on the CAF and the GAC have similar BET surface areas and narrow micropore (<10 Å) volumes. However, the carbon on the CAF has higher pore volume in the pore size range between 10 and 50 Å, and higher H/C atomic ratio in the pore surface. Atrazine adsorption isotherms show that the CAF has a higher adsorption capacity than the GAC, suggesting that higher pore volumes in the pore size range between 10 and 50 Å and higher H/C ratio are advantageous for improved atrazine adsorption from water. Breakthrough curves for both the CAF filter and a commercially available granular activated carbon GAC filter containing equal weight or equal bed volume of adsorbent show that the CAF filter has greatly improved effectiveness of adsorption over the GAC filter for removing trace atrazine below the USEPA standard 3 ppb. Such a CAF filter also displays a much higher effectiveness of selective adsorption of atrazine over the GAC filter in the presence of a 50 times higher concentration of humic acid. The CAF filter could be regenerated to at least 90% of its original activity by heating at 350 °C in air.

Acknowledgements

This work was supported by The WaterCAMPWS, a Science and Technology Center of Advanced Materials for the Purification of Water with Systems under the National Science Foundation agreement number CTS-0120978. Atrazine analysis was carried out at Illinois Waste Management and Research Center (WMRC). We thank CRANE & CO. for the free samples.

References

1. A. L. Aspelin, U.S. Environmental Protection Agency. Pesticides Industry Sales and Usage: 1996 and 1997 Market Estimates, Office of Prevention, Pesticides and Toxic Substances, Washington, DC, 1999.
2. U.S. Environmental Protection Agency. Revised Preliminary Human Health Risk Assessment for Atrazine, Office of Pesticide Programs, Washington, DC, 2002.
3. J. E. Barbash, G. P. Thelin, D. W. Kolpin and R. J. Gilliom, U.S. Geological Survey. Distribution of Major Herbicides in Ground Water of the United States, U.S. Geological Survey, Reston, VA, 1999, 98-4245.
4. International Agency for Research on Cancer, Monographs on the Evaluation of Carcinogenic Risks to Humans: Atrazine, World Health Organization, Geneva, 1999, vol. 73, p. 59.
5. M. A. Kettles, S. R. Browning, T. S. Prince and S. W. Horstman, Environ. Health Perspect., 1997, 105, 1222–1227.
6. J. Kniewald, M. Jakominic, A. Tomljenovic, B. Simic, P. Romac, D. Vranesic and Z. Kniewald, J. Appl. Toxicol., 2000, 20, 61–68.
7. T. E. Stoker, C. L. Robinette and R. L. Cooper, Toxicol. Sci., 1999, 52, 68–79.
8. R. L. Cooper, T. E Stoker, L. Tyrey, J. M. Goldman and W. K. McElroy, Toxicol. Sci., 2000, 53, 297–307.
9. J. Hetrick, R. Parker, R. Pisigan, Jr. and N. Thurman, Progress report on estimating pesticide concentrations in drinking water and assessing water treatment effects on pesticide removal and transformation: A consultation, USEPA, Office of Pesticide Programs, Washington, DC, 2000, p. 28, http://www.epa.gov/scipoly/sap/meetings/2000/september/sept00_sap_dw_0907.pdf.
10. J. Cromphout and W. Rogge, Water Sci. Technol.: Water Supply, 2002, 2(1), 233–40.
11. C. Pelekani and V. L. Snoeyink, Carbon, 2000, 38(10), 1423–36.
12. I. Martin-Gullon and R. Font, Water Res., 2001, 35(2), 516–20.
13. Q. Li, V. L. Snoeyink, B. J. Marinas and C. Campos, Water Res., 2003, 37(20), 4863–72.
14. G. Newcombe, M. Drikas and R. Hayes, Water Res., 1997, 31(5), 1065–73.
15. T. Lebeau, C. Lelievre, D. Wolbert, A. Laplanche, M. Prados and P. Cote, Water Res., 1999, 33(7), 1695–705.
16. Q. Li, V. L. Snoeyink, B. J. Marinas and C. Campos, Water Res., 2003, 37, 773–84.
17. J. Economy, C. L. Mangun and Z. Yue, US Pat. 6517906, 2001.
18. Z. Yue, C. L. Mangun and J. Economy, Carbon, 2002, 40, 1181–91.
19. Z. Yue, C. L. Mangun and J. Economy, Carbon, 2004, 42, 1973–82.
20. Z. Yue, J. Economy and C. L. Mangun, Carbon, 2003, 41, 1809–17.
21. Z. Yue, K. R. Benak, J. W. Wang, C. L. Mangun and J. Economy, J. Mater. Chem., 2005, 15, 3142.
22. Z. Yue, C. L. Mangun, J. Economy, P. Kemme, D. Cropek and S. Maloney, Environ. Sci. Eng., 2001, 35, 2844–8.
23. Z. Yue and J. Economy, J. Nanopart. Res., 2005, 7(4), 477–87.
24. Q. Li, V. L. Snoeyink, C. Campos and B. J. Marinas, Environ. Sci. Technol., 2002, 36, 1510–15.

---