Liquid-Phase Adsorption of Organic Compounds by Granular

Apr 15, 1995 - Liquid-phase adsorption of organic compounds by granular activated carbon (GAC) and activated carbon fibers (ACFs) is investigated. Ace...
8 downloads 0 Views 1MB Size
2110

Znd. Eng. Chem. Res. 1996,34, 2110-2116

Liquid-Phase Adsorption of Organic Compounds by Granular Activated Carbon and Activated Carbon Fibers Sheng H. Lin* and Feng M. Hsu Department of Chemical Engineering, Y u a n Ze Institute of Technology, Neili, Taoyuan 320, Taiwan, ROC

Liquid-phase adsorption of organic compounds by granular activated carbon (GAC) and activated carbon fibers (ACFs) is investigated. Acetone, isopropyl alcohol (IPA), phenol, and tetrahydrofuran (THF) were employed a s the model compounds for the present study. It is observed from the experimental results t h a t adsorption of organic compounds by GAC and ACF is influenced by the BET (Brunauer-Emmett-Teller) surface area of adsorbent and the molecular weight, polarity, and solubility of the adsorbate, The adsorption characteristics of GAC and ACFs were found to differ rather significantly. I n terms of the adsorption capacity of organic compounds, the time to reach equilibrium adsorption, and the time for complete desorption, ACFs have been observed to be considerably better than GAC. For the organic compounds tested here, the GAC adsorptions were shown to be represented well by the Langmuir isotherm while the ACF adsorption could be adequately described by the Langmuir or the Freundlich isotherm. Column adsorption tests indicated that the exhausted ACFs can be effectively regenerated by static in situ thermal desorption a t 150 "C, but the same regeneration conditions do not do as well for the exhausted GAC.

Introduction Liquid organic compounds have been extensively used as important ingredients, and many of them have been used as solvents in a wide variety of manufacturing processes. Hence these organic compounds are commonly present in an industrial manufacturing environment. The environmental problems associated with these organic compounds stem mainly from the presence of these compounds in wastewater in low concentration. A large number of these organic compounds have serious health consequences and some of them are cancer-causing. Hence removal of these organic compounds has become increasingly important and in many cases urgently needed. Biological wastewater treatment method in conjunction with chemical coagulation is by far the most widely used method for dealing with various chemical wastewaters (Reynolds, 1982). This method however is handicapped by the fact that it is only able to treat the wastewater containing relatively low concentration of organic compounds. At higher concentration, a large number of these organic compounds are known to have inhibitory effects on biological growth (McKinney, 1962). Hence other nonbiological treatment methods are needed to overcome this difficulty. Adsorption of organic compounds by activated carbon has been widely considered as a tertiary treatment for final polishing of treated wastewater to meet more stringent discharge requirements than those achieved by the secondary treatment (Reynolds, 1982; No11 et al., 1992; Stenzel, 1993). Hence it is seldom practiced in the industries as a treatment unit primarily because of its relatively higher cost. However, the activated carbon adsorption of volatile or nonvolatile organic compounds has been suggested as a highly viable alternative because of its ability and high efficiency to recover those organic compounds (No11 et al., 1992; Stenzel, 1993). Hence, investigations of the adsorption of organic

* Address correspondence to this author. FAX: 9373.

886-3-455-

compounds by activated carbon have received a considerable amount of attention in the past several decades (Giusti et al., 1974; McKay and Bino, 1990; Ying and Woehr, 1990; Heilshorn, 1991; No11 et al., 1992; Stenzel, 1993; Graham and Ramaratnam, 1993; Chatzopoulos et al., 1994; Sorial et al., 1994). One important aspect regarding the activated carbon adsorption of organic compounds which, until recently, has not received much attention was the organic compound adsorption by activated carbon fibers (Lin and Economy, 1973;Economy and Lin, 1976; Kasaoka et al., 1989a,b; Foster et al., 1992; Cal et al., 1994). Activated carbon fibers (ACF) are essentially activated carbon in a different form. In contrast t o GAC, ACF can be made of polyacrylonitrile, rayon, or pitch (Lin and Economy, 1973; Kasaoka et al., 1989b). Polyacrylonitrile has been the most popular because of relatively good product yield and stability. The raw material is carbonized first at a high temperature in the absence of oxygen and then activated, much like the way GAC is manufactured (Lin and Economy, 1973; Kasaoka et al., 1989a). Although relatively more expensive when compared to GAC, ACF has many distinct advantages, as 'will be elaborated later, and hence is worth looking into for liquid organic compound adsorption. Economy and Lin (1976) appeared to be the first researchers investigating the ACF adsorption of benzene and phenol. These authors only examined the relatively simple adsorption characteristics. Kasaoka et al. (198913) employed ACFs for liquid-phase adsorption of various dyestuffs. Foster et al. (1992) investigated the gasphase ACF adsorption of n-butane, acetone, and benzene. Cal et al. (1994) determined the isotherm parameters for gas-phase adsorption of acetone and benzene. There is little information available in the literature regarding the liquid-phase ACF adsorption of organic compounds. This problem is of practical importance due to the fact that the organic compounds are present in low concentration in many wastewaters from chemical and petrochemical industries. Furthermore, contaminated groundwater near a heavy petrochemical complex may contain many organic compounds in low concentra-

0888-588519512634-2110$09.00/0 0 1995 American Chemical Society

Ind. Eng. Chem. Res., Vol. 34,No. 6, 1995 2111

- -

Table 1. Properties of Three ACF Grades

thickness, cm weight, g/cm2 BET surface area, m2/g pore diameter, A

ACF-900

ACF-1300

ACF-1500

0.065 120 850-950 7-9

0.061 80 1250-1350 8-10

0.057 70 1450-1550 8-10

tion. Air stripping has been a popular method for dealing with these chemical wastewater or contaminated groundwater (Hand et aZ., 1986; Ball and Edwards, 1992). This method is efficient in transferring the low-concentration organic compounds from the liquid phase to the gas phase (Hand et al., 1986; Ball and Edwards, 1992). The offgas from the air stripper however often contains a sufficient amount of organic compounds which exceeds the safe discharge requirement. Therefore removal of organic compounds from the offgas is becoming necessary and adsorption by granular activated carbon has been proposed for this purpose (Critenden et al., 1988). The combination of air stripping and activated carbon adsorption seems cumbersome. Direct activated carbon adsorption of organic contaminants in the chemical wastewaters or groundwater appears to be more straightforward and economical. The purpose of the study focuses on investigating the liquid-phase-adsorption characteristics by granular activated carbon (GAC) and activated carbon fiber (ACF). Experiments were conducted to gather data which permit detailed analysis and evaluation of the GAC and ACF adsorption characteristics on a common basis.

* *

a

1. GAC

2. ACF-900

3. ACF-1300

Experimental Studies The GAC employed here, as supplied by Chang Chun Chemical Co., Taiwan, was of industrial grade and made from coconut shells. According to the manufacturer, the GAC had an average granule diameter of 5 mm, a density of 0.45 g/cm2, a BET surface ?rea of 950 m2/g, and an average pore diameter of 26 A. This grade of GAC was widely used in the chemical and food industries for adsorption, decolorization, and pollution abatement purposes. The activated carbon fiber (ACF) was manufactured by Yih Hsin Chemical Co., Taipei, Taiwan. It was made of polyacrylonitrile. Three grades of ACF were tested here. The properties of the three ACFs are listed in Table 1. Figure 1-4 shows the micrographs of GAC and three ACFs enlarged 15 times by a microscope (Model CHT, Olympus Optical Co., Tokyo, Japan). Figure 1-1 clearly displays the irregular GAC geometry. The dark portion in Figure 1-2 to 1-4 is the entwined activated carbon fibers whereas the white portion is the interstitial spaces among the fibers. It is apparent that there is an increasing degree of regularity in the shape of the interstitial rectangles as the BET surface area of ACF increases. This is due to the denser chemical fibers used for the ACF-1500 manufacturing. The organic compounds chosen for the present study were acetone, isopropyl alcohol (IPA), tetrahydrofuran (THF), and phenol (all GR grade solvents supplied by E. Merck GmbH, Darmstadt, Germany). An aqueous solution with a 400 mg/L concentration was prepared by mixing an appropriate amount of solvent with deionized water. In the experiments, 150 cm3 of the prepared solution was placed in a 200 cm3 flask. Various known quantities (in the range between 1.5 and 2.5

4. ACF-1500

Figure 1. Micrographs (magnified 15 times; reproduced a t 65%. of original size) of GAC and ACFs.

g) of pretreated GAC or ACF were added to the flasks which were then sealed. About 10 flasks were prepared in the same fashion for each experimental run. The flasks were put in the constant-temperature bath maintaining at 20 "C. The shaker speed of the constant temperature bath was set a t 80 cycles/min to simulate the mixing condition. At a desired time, a flask was removed from the constant-temperature bath and the liquid sample was taken for measuring the concentration of organic compound using a HP gas chromatograph (GC; Model 5890, Hewlett Packard Co., Denver, CO) with a flame ionization detector and GP 80/100 Carbopack packed column. The time-dependent concentration changes of the organic compounds allowed determination of the GAC and ACF adsorption capacities and calculation of the isotherm parameters. A GAC or ACF column adsorption apparatus, as shown in Figure 2, was also built in the present study to examine the GAC or ACF column adsorption process. The apparatus mainly consisted of a Pyrex column 3.1 cm in diameter (D) and 20.5 cm long (L). The GAC was

2112 Ind. Eng. Chem. Res., Vol. 34, No. 6, 1995

1

A

Rotameter

8-1,.

Temperature Controller

Packe

I

'

I

Feed Pump

Reservoir

HP G C Figure 2. Apparatus of GAC and ACF-1300 column adsorptions.

randomly packed in the column while the ACF was tightly wrapped in a cylindrical form which was then inserted into the column. The LID ratio of the column was sufficient to minimize the end effect of the adsorption column. The column was surrounded with heating tape for temperature control. A solution containing the organic compound at a concentration of 100 mg/L was prepared. During operation, the solution was pumped to the top of the adsorption column at a flow rate of 10 mumin. The concentration of organic compound at the exit of the column was monitored by the HP 5890 gas chromatograph. After the adsorption column was completely exhausted, the heater was set at 150 "C with an attempt to desorb the adsorbed organic compound. The static thermal desorption process lasted 24 h. After the desorption was finished, a second adsorption run was performed.

Discussion of Results Acetone rate of adsorption by GAC at 20 "C is shown in Figure 3 for two different mixing conditions. It is clear in this figure that the mixing does not affect the equilibrium adsorption capacity which is about 180 mg of acetonelg of GAC for both mixing and unmixing cases. The figure also reveals that without mixing, the GAC adsorption took about 200 min to reach equilibrium while with mixing the equilibrium adsorption time was shortened by one-fourth. The results are not unexpected because of reduced mass transfer resistance under mixing condition. The corresponding acetone adsorption by ACF-1300 is demonstrated in Figure 4. The time to reach equilibrium becomes approximately 180 min for both cases with and without mixing. Comparison of Figures 3 and 4 also reveals that the adsorption capacity of ACF-1300 exceeds that of GAC by more than two-thirds, which represents a significant margin. The BET surface area of adsorbent is an important factor affecting the adsorption capacity of the adsorbents (No11 et al., 1992). Hence it would be of interest to compare the equilibrium adsorption capacities of the

200

u

I

v

// n

I

I

4

II;/ 0

/

I

/

/

A with 80 cycle/min mixing 60

120

180

240

300

Time, min Figure 3. Effect of mixing on the GAC adsorption rate of acetone a t 20 "C.

various absorbents tested in the present study. Table 2 compares the adsorption capacities of adsorbents for acetone, IPA, THF, and phenol. It is apparent that the effect of BET surface area on the adsorption capacity is quite significant. Such an effect is even more pronounced for the three ACFs. However, it should be noted that the increasingly superior adsorption capacity of ACFs is accompanied with relatively higher unit cost of the adsorbate. Adsorption of organic compounds by adsorbents has been known to be affected by the polarities of adsorbent and adsorbate and the solubility of adsorbate in solution (Adsomson, 1988). In general, nonpolar organic compounds are preferentially adsorbed by nonpolar adsorbent from polar solvent (water) while polar organic compounds are preferentially adsorbed by polar adsor-

Ind. Eng. Chem. Res., Vol. 34,No. 6, 1995 2113 300

250

-

200

-

-

150 A

0

A 0

Without mixing

With 80 cycle/min mixing 60

120

180

240

300

0

60

Time, min Figure 4. Effect of mixing on the ACF-1300 adsorption rate of acetone. Table 2. Equilibrium Adsorption Capacities of GAC and ACFs adsorption adsorbent adsorbate capacity, mg/g 179 GAC acetone IPA 135 203 THF 138 phenol 215 ACF-900 acetone IPA 171 221 THF 187 phenol 296 ACF-1300 acetone IPA 248 THF 346 271 phenol 358 ACF-1500 acetone 305 IPA THF 387 325 phenol

bents from nonpolar solvents (Adamson, 1988). In addition, the molecular weight of organic chemical compounds plays a very strong role in the adsorption process. In the present case, acetone has the largest polarity with a dipole moment of 2.88 (Dean, 19871,but has the lowest molecular weight (58) among the four organic compounds. On the other hand, phenol has the lowest polarity with a dipole moment of 1.44, but has the largest molecular weight. THF comes second in both molecular weight and polarity. Among the four organic chemical compounds, acetone and IPA are miscible with water in any proportion while phenol and THF are less soluble in water. Figures 5 and 6, respectively, display the GAC and ACF-1300 adsorption rates of the four adsorbates. The time to approach equilibrium significantly varies with species. The adsorption capacity of both adsorbents generally follows the order THF > acetone > phenol> IPA. Such order of adsorption capacity clearly demonstrates the strongly interacting and complex phenomena of molecular weight, polarity and solubility among these four organic compounds.

120

180

240

300

Time, min Figure 5. GAC adsorption rates of various organic compounds.

i

350

150

100

0

60

120

iao

240

300

Time, min Figure 6. ACF-1300 adsorption rates of various organic compounds.

The two general monolayer adsorption iostherms can be represented by the well-known Langmuir and Freundlich models as

Q = Kcelin (2) where Q is the equilibrium adsorption capacity (mg of adsorbatelg of adsorbent), C, is the concentration of organic compound at equilibrium ( m a of adsorbate), and A, B , K and lln are constant parameters to be determined. The equilibrium adsorbate concentration (C,) was measured by GC after the adsorption equilibrium was reached. With known initial adsorbate con-

2114 Ind. Eng. Chem. Res., Vol. 34, No. 6, 1995

E 1

3001

400

6.0

/'

200

I

I

-1

5.0

cp.5*4I+

F:

5.2

-

5.0

-

4.0

-

Freundlich

4.61

3.0 0

20

40

a0

60

100

Ce , mg/l Figure 7. Equilibrium GAC adsorption of acetone as a function of equilibrium acetone concentration (C,).

t

,

I

3.1

, 3.LT ,3f03, I

I

3 & .,

Q)

8

150

Table 3. Parameters for the Langmuir and Freundlich Isotherms adsorbent

adsorbate

A

ACF-1300

acetone IPA THF phenol acetone IPA THF Phenol

408.82 372.75 449.82 464.57 393.76 369.76 407.29 407.48

GAC

c, Q)

2 bD

/ P

100

/

\

bn

E

n

a

0

I

adsorbent GAC

0

ACF-1300

I

I

50

100

I

150

B 0.0273 0.0133 0.0520 0.0118 0.0039 0.0022 0.0046 0.0018

Freundlich

K 25.61 12.59 43.68 11.01

lln

0.5501 0.6112 0.5218 0.6891

Table 4. Parameters for the Temperature-Dependent Equilibrium Adsorption Correlation

Freundlich

J

50

3 B

Figure 9. Equilibrium GAC and ACF-1300 adsorption capacities of acetone as a function of temperature.

Langmuir 200

1

3.5

I

I

200

I

250

Ce , mg/l Figure 8. Equilibrium ACF-1300 adsorption of acetone as a function of equilibrium acetone concentration (C, 1.

centration (CO)and the amount of adsorbent, the equilibrium adsorption capacity of the adsorbate (Q) was determined. Using the equlibrium experimental data (Q and Cd, the isotherm parameters in eq (1)and 2 can be, respectively, obtained by plotting CJQ against 1/C, according to eq (1)and In(&) against ln(C,) according to eq (2). Figures 7 and 8 show comparisons of the two computed isotherms by eqs (1)and 2 using the parameters obtained and the measured data. For GAC adsorption of acetone, the Langmuir isotherm appears to consistently fit the data better than the Freundlich isotherm. On the other hand, both isotherms seem to fit the measured data of ACF adsorption of acetone equally well. The same phenomena were also observed for the GAC and ACF-1300 adsorptions of other organic

adsorbate

KO

acetone THF IPA phenol acetone THF IPA phenol

6.4873 3.0727 10.5528 49.0216 3.2579 2.5796 3.7286 74.9716

D 976.12 1122.05 864.01 297.53 718.26 864.01 625.14 376.37

compounds. The parameters obtained for the GAC and ACF-1300 adsorption isotherms are listed in Table 3. The data discussed above were obtained from experiments conducted at 20 "C. Temperature is known to play an important role in the activated carbon adsorption. Hence it would be of practical interest t o examine the temperature effect on the present adsorption process. Figure 9 shows such a temperature effect. The semilogarithmic plot of the adsorption capacity vs 1/T appears t o correlate very well indeed. Such a relation can be represented by the following exponential equation

Q = KOexp(DIT)

(3)

in which KOand D are constant parameters and T is the temperature (kelvin). The parameters estimated from the measured data are listed in Table 4 for GAC and ACF-1300 adsorptions of acetone, IPA, THF, and phenol. Figures 10 and 11 show, respectively, the first and second GAC and ACF-1300 column adsorptions of

Ind. Eng. Chem. Res., Vol. 34, No. 6, 1995 2115 1.o

0.8

*

0.4

-/ 0.2 -

0.0 +

o

i

2

3

5

4

6

7

a

9

io

0.0

Time, hr

*

0.4

-

i 1 I 4 / +

1 0

1

2

3

4

5

6

6.0

8.0

10.0

Figure 12. Comparison of breakthrough curves of GAC and ACF1300 adsorptions of IPA as a function of time with 10 m u m i n liquid flow rate and 100 mgA inlet acetone concentration.

Conclusions

lst

I

0.2

4.0

first and second GAC adsorptions of acetone in Figure 10. In Figure 11, the breakthrough points for both ACF adsorptions arrive a t 2.8 h, implying a much improved ACF adsorption performance than GAC. Figure 12 compares the column GAC and ACF adsorptions of acetone. The adsorption characteristics are seen to be similar to those observed in Figures 10 and 11,confirming better ACF adsorption characteristics than the GAC counterpart.

-

I

2.0

1

Time, hr

Figure 10. Breakthrough curves of the first and second GAC adsorptions of acetone as a function of time with 10 m u m i n and 100 mg/L inlet acetone Concentration.

0.a

1

I

7

8

9

10

Time, hr Figure 11. Breakthrough curves of the first and second ACF1300 adsorptions of acetone as a function of time with 10 m u m i n and 100 mg/L inlet acetone concentration.

acetone. Note that the first adsorption was that using virgin GAC or ACF-1300 while the second adsorption was that using regenerated GAC or ACF-1300. In these figures, COis the initial concentration of organic compound and C the concentration at various adsorption times. It is apparent in Figure 10 that there is a significant difference between the two breakthrough curves of GAC adsorption, which represent the first and second adsorptions, respectively. But Figure 11reveals very little difference between the two breakthrough curves of ACF-1300 adsorption. Taking 0.05 of C/Coas the breakthrough point (Reynolds, 19821, the times to reach this point are 2.2 and 1.6 h, respectively, for the

Experimental results of batch and column adsorption tests in the present study show that organic compound adsorption by GAC and ACFs is strongly influenced by the BET surface area of adsorbent and the molecular weight, polarity and solubility of the adsorbate. Experimental results indicate that ACF adsorptions significantly outperform GAC adsorption under the same operating conditions. Both GAC and ACF adsorptions are highly temperature dependent. The adsorption capacity is found to correlate well t o the reciprocal of temperature by an exponential equation. The equilibrium adsorption data also show that GAC adsorption of organic compounds can be represented by the Langmuir isotherm. However the ACF adsorption can be reasonably well described by either the Langmuir or Freundlich isotherm. The column adsorption tests reveal much earlier arrival of the GAC breakthrough point than the ACFs, implying much improved ACF adsorption performances. Static in situ thermal desorption experiments at 150 "C of the exhausted GAC and ACFs were also conducted. The results shows that ACFs could be essentially completely regenerated at t h s temperature, but regeneration of exhausted GAC is far from complete at the same temperature.

Acknowledgment The authors sincerely thank the National Science Council, Taiwan, ROC (under Grant NSC84-2621-P1550011, for the financial support of this project.

2116 Ind. Eng. Chem. Res., Vol. 34, No. 6, 1995

Literature Cited Adamson, A. W. Physical Chemistry of Surfaces, 2nd ed.; Interscience Publishers: New York, 1988. Ball, B. R.; Edwards, M. D. Air Stripping VOCs from Groundwater: Process Design and Considerations. Environ. Prog. 1992, 1 1 , 39. Cal, M. P.; Larson, S.M.; Rood, M. J. Experimental and Modeled Results Describing the Adsorption of Acetone and Benzene onto Activated Carbon Fibers. Environ. Prog. 1994,13, 26. Chatzopoulos, D.; Varma, A.; Imine, R. L. Adsorption and Desorption Studies in the Aqueous Phase for the ToluendActivated Carbon System. Environ. Prog. 1994,13, 21. Crittenden, J. C.; Cortright, R. D.; Rick, B.; Tang, S. R.; Parson, D. Using Granular Activated Carbon to Remove VOCs from Air Stripper Off-Gas. J . Am. Water Works Assoc. May, 73. Dean, J. H. Lunge’s Handbook of Chemistry, 17th ed.; McGrawHill: New York, 1987. Economy, J. and Lin, R. Y. Adsorption Characteristics of Activated Carbon Fibers. Appl. Polym. Symp. 1976,29,199. Foster, K. L.; Fuerman, R. G.; Economy, J.; Larson, S. M.; Rood, M. J. Adsorption Characteristics of Trace Volatile Organic Compounds in Gas Streams onto Activated Carbon Fibers. Chem. Muter. 1992,4,1068. Giusti, D. U.; Conway, R. A,; Lawson, C. T. Activated Carbon Adsorption of Petrochemicals. Chem. Eng. 1974,46, 947. Graham, J. R.; Ramaratnam, M. Recover VOCs using Activated Carbon. Chem. Eng. 1993,45 (31, 58. Hand, D. W.; Critenden, J. C., Gehin, J. L.; Lykins, B. W. Design and Evaluation of a n Air Stripping Tower for Removing VOCs from Grooundwater. J . AWWA 1986,Sept, 87. Heilshorn, E. D. Removing VOCs from Contaminated Water. Chem. Eng. 1991,43 (61, 120. Kasaoka, S.; Sakata, Y.; Tanaka, E.; Naitoh, R. Design of Molec-

ular-Sieve Carbon: Studies of the Adsorption of Various Dyes in the Liquid Phase. Znt. Chem. Eng. 1989a,29,102. Kasaoka, S.; Sakata, Y.; Tanaka, E.; Naitoh, R. Preparation of Activated Fibrous Carbon from Phenolic Fabric and its Molecular-Sieve Properties. Znt. Chem. Eng. 1989b,29,101. Lin, R. Y.; Economy, J. The Preparation and Properties of Activated Carbon Fibers Derivated from Phenolic Precursor. Appl. Polym. Symp. 1973,21,143. McKay, G.; Bino, M. J. Fixed Bed Adsorption for the Removal of Pollutants from Water. Environ. Pollut. 1990,66, 32. McKinney, R. E. Microbiology for Sanitary Engineers; McGrawHill: New York, 1962. Noll, K. E.; Gournaris, V.; Hou, W. S.Adsorption Technology for Air and Water Pollution Control; Lewis Publisher: Chelsea, MI, 1992. Reynolds, T. D. Unit Operations and Processes in Environmental Engineering; Wadsworth, Inc.: Belmont, CA, 1982. Sorial, G. A,; Suidan, M. T.; Vidic, R. D.; Brenner, R. C. Effect of GAC Characteristics on Adsorption of Organic Pollutants. Enuiron. Prog. 1994,13, 53. Stenzel, M. H. Remove Organics by Activated Carbon. Chem. Eng. Prog. 1993,89 (41, 36. Ying, W. C.; Woehr, G. C. Adsorption Capacities of Activated Carbon for Organic Constituents of Wastewater. Enuiron. Prog. 1990,9,120.

Received for review September 22, 1994 Revised manuscript received February 15, 1995 Accepted March 13, 1995@ I39405596 Abstract published in Advance ACS Abstracts, April 15, 1995. @