Adsorption of large compounds on activated carbon

MISSING IMAGE

Material Information

Title:
Adsorption of large compounds on activated carbon
Series Title:
Florida Water Resources Research Center Publication Number 74
Physical Description:
Book
Creator:
Arbuckle, W. B.
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Notes

Abstract:
As Florida's water demand increases due to population growth and to increased per capita consumption, less desirable sources of water will be used, requiring additional treatment for the removal of organic compounds. In addition, wastewaters will require additional treatment to remove organic compounds. In both cases, activated carbon adsorption is the process most likely to be applied. Most of the organics present in natural waters and biologically treated wastewaters.are large compounds (i.e. molecular weight MW > 500), and little is known about large compound adsorption. An optimum size of adsorbate molecule appears to exist, since the 5200 MW inulin adsorbs better than the 342 MW sucrose and the 20,000 MW xylan on all three adsorbents evaluated, including the small-pored petroleum pitch carbon. When adsorbing the xylan, even the large-pored lignite carbon's capacity was much reduced. No chromatographic effect was noted in continuous column studies at 1 and 2 gpm/ft2. Premature exhaustion may exist as proposed in previous studies, but additional studies are needed at longer contact times (> 30 minutes); in general, carbon use decreases with increased contact time. More studies are needed.

Record Information

Source Institution:
University of Florida Institutional Repository
Holding Location:
University of Florida
Rights Management:
All rights reserved by the source institution and holding location.
System ID:
AA00001553:00001


This item is only available as the following downloads:


Full Text
















Publication No. 74


ADSORPTION OF LARGE COMPOUNDS ON ACTIVATED CARBON


By


W. B. Arbuckle


Department of Environmental Engineering Sciences
University of Florida
Gainesville


.-

i

















ACKNOWLEDGEMENTS

The research support of the Florida Water Resources Research

Center under the direction of Dr. James P. Heaney with funds

provided by the U. S. Department of the Interior through their

annual cooperative program .(OWRT Project No. A-045-FLA) and the

Engineering and Industrial Experiment Station of the University

of Florida, Gainesville, FL. is greatly acknowledged. In

addition, Amy Jo Nelson's laboratory work on the adsorption

equilibria and her master's degree special project report

(reference 0) are acknowledged, as is the laboratory work of Jon

Earle on the column studies.














TABLE OF CONTENTS







Acknowledgements .......... .............. .... ...... i

List of Tables............... .......................iii

List of Figures........... ...................iii

Abstract.. .. ...... .. .. ...... iv

Introduction ... .......................... 1

Background..................... ...... ...... .......... 4

Humic substances... ... .......................L.. 5

Wastewater materials................. ............ 7

Pure compounds ............ ..................... 10

Experimental. ................... ...................... 13

Materials.......... ... .............. ........... 13

Equilibria studies.............................. 14

Column studies.............................. .. 15

Analytical methods.......................... 17

Results and Discussion....................... ........ 18

Equilibria tests....................... 1.....8.

Column tests................. .................... 31

Conclusions ............................................ 40

References ........................ .................. 42





ii












LIST OF TABLES


Table I Pore Volume Distribution for Adsorbents...... 14

Table II Freundlich Constants........................ 19

Table III Times to Breakpoint 1 gpm/ft2 system....... 35

Table IV Times to Breakpoint 2 gpm/ft2 system....... 38





LIST OF FIGURES


Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure


Figure 11


Figure 12


Figure 13


Figure 14


- Modified Shaker Table for Isotherms......... 16

- Roller Apparatus for Isotherms ............ 16

- Sucrose Isotherms.......................... 21

- Inulin Isotherms............................. 23

- Xylan Isotherms............................. 24

- Isotherms on BAC-SQ........................ 26

- Isotherms on F400 ......... ........ ........ 27

- Isotherms on HD4000 ........................ 29

- Miscellaneous Compounds on F400............. 30

- Breakthrough Curves for 1 gpm/ft2, PEG 1000
System...................................... 34

- Breakthrough Curves for 1 gpm/ft2, PEG 14,000
System...................................... 34

- Breakthrough Curves for 2 gpm/ft2, PEG 1000
System..................................... 37

- Breakthrough Curves for 2 gpm/ft2, PEG 14,000
System .............. ........................ 37

- Carbon Usage to a 30 mg/L Breakpoint........ 39












ABSTRACT


As Florida's water demand increases due to population growth

and to increased per capital consumption, less desirable sources

of water will be used, requiring additional treatment for the

removal of organic compounds. In addition, wastewaters will

require additional treatment to remove organic compounds. In

both cases, activated carbon adsorption is the process most

likely to be applied. Most of the organic present in natural

waters and biologically treated wastewaters are large compounds

(i.e. molecular weight [MW] > 500), and little is known about

large compound adsorption.

An optimum size of adsorbate molecule appears to exist,

since the 5200 MW inulin adsorbs better than the 342 MW sucrose

and the 20,000 MW xylan on all three adsorbents evaluated,

including the small-pored petroleum pitch carbon. When adsorbing

the xylan, even the large-pored lignite carbon's capacity was

much reduced. No chromatographic effect was noted in continuous

column studies at 1 and 2 gpm/ft2. Premature exhaustion may

exist as proposed in previous studies, but additional studies are

needed at longer contact times (> 30 minutes); in general, carbon

use decreases with increased contact time. More studies are

needed.







INTRODUCTION

Water and wastewater problems are likely to become very

important issues in Florida in the near future; due to its

rapidly expanding population. The problem is intensified since

most of the population growth is occurring in the southern

coastal regions of the state where the least potable water is

available. Groundwater is used extensively for water supply, but

additional water withdrawals may threaten existing supplies by

permitting salt water to intrude into the fresh water aquifers,

rendering them useless. Due to the large number of hazardous

waste disposal sites within the state, the potential for

contaminating the groundwater is great. And some water supply

wells have been abandoned due to pollution from the surface.

Groundwater supplies are limited, and diligence is needed to

protect the existing systems.

Many communities use surface waters to supply their water;

surface water sources could be used more extensively in Florida,

but water quality problems are likely to result, as well as legal

problems associated with transferring water from one drainage

basin to another, since most of the surface water is in the

northern part of the state. Many surface waters are highly

colored due to decaying vegetation, the color is not only

aesthetically objectionable, but produces trihalomethanes, which

are believed to be carcinogenic compounds. Special treatment

will be required to remove this color, with both activated carbon

adsorption and chemical coagulation the most likely processes.

In addition, surface water use will require protection from

wastewater discharges, especially as the population expands.









More extensive wastewater treatment will be required when the

waste is discharged into a stream used for potable water supply

or when the waste is used to recharge the groundwater, whether to

protect it from salt water intrusion or as an ultimate disposal

method. Undoubtedly, concern for trace organic compounds in

these wastes will require activated carbon adsorption as a

treatment step.

Activated carbon adsorption is a very effective process for

removing organic compounds from water. In potable water

treatment, activated carbon has been used for years to remove

taste and odor causing compounds and for removing color

bodies (1). When used in these applications, powdered activated

carbon is usually added in the normal water treatment processing

steps; if the water has extremely high levels of contaminants,

then granular activated carbon has been used to replace the sand

in the rapid sand filters (2) -- providing better, more

economical treatment. Recently adsorption has been extensively

evaluated for the removal of trihalomethane precursors and

synthetic organic chemicals from our potable waters (3,4,5), with

the U. S. Environmental Protection Agency proposing it as a

treatment requirement for all water plants providing water to

communities greater than 75,000 people (6). For these

applications, an additional processing step would probably be

added to the treatment system, rather than replacing the sand with

activated carbon as is done for high levels of taste and odor

producing compounds and for high color levels, adding substantially

to the capital and operating costs of the water treatment system (7).







When a groundwater becomes contaminated by synthetic chemicals

and the water source is not abandoned, the water can be processed

to acceptable quality using granular activated carbon systems

(8). Portable systems can be taken to the site and installed in

a short time, providing excellent removal of many compounds (9).

In wastewater treatment, activated carbon adsorption is used

to remove refractory organic compounds when a high quality water

is needed such as for South Lake Tahoe (10) or Water Factory

21 (11). It is also used with chemical addition to replace

"normal" secondary treatment plants with physical/chemical

treatment processes (12), where the adsorption system is used to

remove the organic compounds. Industry can also use activated

carbon to remove toxic and nondegradable compounds before

discharging them into sewers (13), so they won't destroy or

inhibit the biological treatment system -- industrial

applications are likely to increase dramatically if the

U.S.E.P.A. sets strick limits on toxic compound discharges.

In many potable water supplies, a variety of organic

compounds exist in the water, with over 250 chemicals identified

in the nation's potable water supplies (14). Extensive efforts

are underway to identify the components of a water, but only a

small fraction of the total organic carbon of a water is

identifiable (15). These compounds are generally small molecules

(smaller than 200 atomic mass units). The remaining compounds

are frequently classified as humic substances which include both

humic and fulvic acids humicc acids are less soluble under acidic

conditions). Their molecular weight distribution has been

evaluated using gel chromatography (16, 17, 18). Most of the









compounds have a molecular weight greater than 500, with many

being considerably larger than 10,000.

In wastewaters, only a small fraction of the organic carbon

present is identifiable as specific compounds -- most of the

remaining compounds are large compounds. Attempts have been made

to identify the families of compounds, with proteins, carbohydrates,

lignins, and tannins sometimes identified along with the molecular

weight distributions.

The purpose of this study is to provide basic background

adsorption equilbria data for some specific high molecular

weight model compounds; this should indicate which families of

compounds could potentially be a problem for adsorption

processes. In addition, how well the model compounds adsorb on

different activated carbons will be evaluated to determine if a

particular carbon appears better suited for large compound

adsorption. Fixed-bed adsorption studies will also be performed

with a mixture of compounds to determine if the compounds

fractionate by molecular weight within the adsorber. [This could

be due to the slower diffusion of the larger compounds, causing

the large compounds to penetrate deeper into the bed where they

adsorb and block the adsorbent's pores; thereby preventing

additional adsorption when the smaller compounds reach that point

in the column. This phenomenon is referred to as premature

exhaustion and results in larger than expected carbon usage (19)].

BACKGROUND

There is not much information available on the adsorption of

large molecular weight compounds, most of the available







information is for humic substances or fractionated wastewater.

In addition, some information is contradictory.

Humic Substances: These materials are very complex,

source-dependent organic substances. They are classified as

humic or fulvic acids, with the humic acids being less soluble

compounds which are separated from the fulvics by lowering the pH

to 2, where the precipitate that forms is humic acid.

The large fraction (MW > 4000) of lake water was found to

adsorb better on aluminum oxide solids than the small fractions

(MW < 400); their adsorption is more pH dependent (20).

Adsorbing fulvic acids extracted from Michigan peat onto

activated carbon finds the large materials (MW > 50,000) adsorbed

less than the smaller ones (MW < 1000). There are indications that

the small fractions associate with the large ones in solution

when adsorbing, since the unfractionated sample's adsorption

pattern was similar to that of the large fraction's (21). The

adsorption of the small compounds correlates with the pore volume

in pores smaller than 7 nanometers for the nine activated carbons

tested, and the large compounds's adsorptive capacity correlates

with the pore volume in pores smaller than 40 nanometers.

Carbons with most of their pore volumes in the larger diameter

pores would be more suitable for adsorbing large compounds. Both

adsorption capacity and adsorption rate increased with decreasing

molecular weight for the fulvics studied. Other researchers have

also found adsorptive capacity to increase with decreasing

molecular weight (22, 23). Unfractionated soil fulvic acid

adsorbed similar to the smallest molecular weight fraction (MW <

5000), while the soil humic acid's adsorption was similar to the









middle molecular weight fraction's (5,000 50,000) (23).

Removals depend upon the source of humic materials (24).

The solution's chemical conditions significantly affect the

results. Humic acid was 10 times more adsorbable from tap water

than from deionized-distilled water, indicating that co- and/or

counterions are important factors (25). Magnesium ions enhanced

adsorption capacity more than calcium; hypochlorite ions also

increased capacity. Others have found little adsorption of

humics in the absence of salts, with the anions having no effects

(26). The cations were most important, with Ca enhanc'.-,

adsorption more than Mg which had a greater effect than Na.

Cation effects were much reduced at low pH values. In another

system phosphate was found to greatly increase adsorption

capacity (2 to 3 times) (23). The humic substances remaining

after precipitation of humic materials with alum, adsorb

substantially better than those present before treatment (27,

21). Solution pH is also important, with studies fin0 '..1:'

adsorption capacity to increase with decreasing pH from 11 to 2

(23) while others found an optimum pH when using an aluminum

oxide adsorbent instead of activated carbon (20). : ere

found to have an optimum adsorption pH in a study testing only 3

pHs, since the lowest pH (4.5) resulted in a lower adsorptive

capacity than the two higher pHs (28).

Humic substances have a significant adverse effect on the

adsorption of chlorophenols (29), while only exhibiting a slight

effect on the adsorption of phenol at neutral pH (30). A larger

adverse effect on phenol adsorption resulted at pH 9, but this is







probably due to the ionization of the phenol as its pKa is

approached (31). At pH 2 ,no interference with adsorption was

observed (30), the humic materials would be expected to adsorb to

a greater extent at pH 2 since they become insoluble; but the

humic materials were believed to become insoluble colloids and

therefore did not adsorb in the normal sense, but behaved in a

manner similar to other colloids (30). Other colloids (clay or

polyelectrolytes) added to the phenol system did not interfere

with phenol adsorption by adsorbing and blocking the pores (30).

Few adsorption column studies were reported. In one study

a short empty-bed contact time was used (1.6 minutes,

approximately equal to a 5 inch deep adsorber with a flow of 2

gpm/ft2) and a "rapid" humic acid breakthrough was observed with

only a small amount of material adsorbed (0.23 mg/g) (28).

Humic substance adsorption has been modeled, with only 40 percent

of the carbon's equilibrium capacity utilized at 80% breakthrough

due to slow adsorption (32). Model studies found the system

insensitive to the film transfer coefficient and, in the early

stages of the breakthrough curve, to the surface diffusion

coefficient; during the later stages of breakthrough, the system

becomes sensitive to surface diffusion. Pretreatment of the

humic substances with alum increases the service time of the

adsorber by 14 to 22 times (32)! Because of the slow surface

diffusion, long empty-bed contact times are required to contain

the mass transfer zone.

Wastewater Materials: Several studies exist using waste-

water compounds. In an early study, compounds larger than 1200

MW were claimed not to adsorb, with only the compounds about 400









MW and smaller adsorbed (33). They hydrolyzed the larger

compounds to smaller ones using lime and a sufficient reaction

time; then the hydrolyzed organic were ,A-'.o.rbed with far better

treatment resulting. Similar results were obtained in another

physical-chemical scheme evaluated at the pilot level (34).

In addition, a study on virus removal found the organic removal

to increase substantially after lime treatment, like in the

hydrolysis treatment (35). Many others have found the smaller

compounds to account for most of the non-adsor'-:'."i material

(36, 37,38, 39). These studies were more detailed research

studies; they all used secondary effluents. The studies showing

the small materials to be more adsorbable used physical/chemical

pilot plant influents and effluents, except for the virus st,.1-

which used a secondary effluent.

In column studies, all MW fractions were f:0. :! to be adsorb-

able to some extent, with 52 percent of the smallest .i ds

and 35 percent of the larger compounds adsorbed; the intermediate

size was 90 percent removed (38). The large ... :-.. .. (MW >

50,000) that were not adsorbed were car'...-y".:ate-like materials.

Column studies performed on size-fractionated biol .;call

treated wastewaters found the smallest :- .*,7 not to adsorb;

but, if the material was fractionated .-:.- ads -.o' .on, all

fractions adsorbed to some extent in the columns (37). Mean cell

residence time greatly affects the quantity of the lar:'

......,,, with decr* .-:-"i mean cell residence times resulting in

more large -. .. ;......:...:, and therefore p.. .rer overall removals by

ads -- ::. on.







Batch kinetic and equilibria studies were performed on three

fractions of an extensively treated industrial wastewater (60

hours hydraulic residence time, sludge age greater than 10 days)

(19). This waste was used in a previous pilot study where adding

a second column in series with the first adsorber resulted in a

70 % increase in carbon usage, rather than the expected reduction

in use (40); it was referred to as premature exhaustion. Pilot

scale ultrafiltration and reverse osmosis was used to fractionate

the waste prior to the adsorption studies. A majority of the

material (48%) passed through the ultrafiltration membrane, but

was rejected by the reverse osmosis membrane (about 342 to 6500

MW). This fraction adsorbed best (79%), while the smallest

material adsorbed least (55%); the large fraction was 62%

adsorbable (19). Adsorption energy increased with increasing

molecular weight, indicating that the larger compounds are more

strongly adsorbed and therefore are more difficult to displace.

The larger compounds adsorb at a slower rate, as would be expected.

This study proposed that the large compounds penetrate far deeper

into the adsorber than the smaller compounds before adsorbing.

Deep in the adsorber, there is less competition for adsorption

sites and they slowly adsorb,but over a period of time they have

utilized a large portion of the adsorbent's capacity. Due to

their size, they effectively block the pores of the carbon and

since they are strongly adsorbed, they cannot be displaced.

Gums were tested in dye wastes for their effects on

adsorption, with the large gums (MW > 200,000) having a positive

effect on the activated carbon system at 65 to 100 mg/L levels,

but having an adverse effect at lower concentrations -- no









reasons were given (41). The effect of pH on the adsorption

of lignosulphate and tannins were evaluated in real industrial

wastes (42), with lignin and tannins both removed better from

their respective wastes at higher pHs.

Pure Compounds: Lysozyme (MW about 13,900) and bovine

serum albumin (MW about 67,500) occupy nearly the same surface

area on activated carbon when adsorbed from solution, even

though one is 4 times larger (43). These proteins were

irreversibly adsorbed. Effective adsorption diffusivities were

about 6 percent of the bulk diffusivity for the lysozyme and 0.7

percent, for the bovine serum albumin. Maximum adsorption

occurred at their isoelectric points.

The equilibria and kinetics of adsorbing a series of

polyethylene glycols (PEGs, MW 194 to 2,490,000) were studied using

coconut-based and lignite-based activated carbons (44), The

coconut carbon has most of its pore volume in micropores (in this

case >70% of the volume is in pores smaller than 3 nanometers),

this does not permit the larger compounds to adsorb as well, with

adsorption capacity decreasing for PEGs larger than 1500 MW.

With the lignite carbon (>70% of its pore volume is in pores

larger than 3 nanometers), adsorption capacity increased with

increasing MW up to 1500 MW, where the maximum loading was

achieved. Increasing MW above this value did not change the

adsorption capacity, so the same surface area or pore volume is

accessible for all these large compounds. The effective diffusion

coefficient decreased with increasing MW, with the coconut carbon's

diffusivities larger than the lignite carbon's. Using these







values in calculations for a "typical" water treatment

application found that compounds much smaller than 10,000 MW

would pass through the adsorber without having a chance to

adsorb, longer contact times are needed.

Three different molecular weights of polyvinyl acetate were

adsorbed on activated carbon; the smaller compounds took longer

to reach "complete" equilibrium; the 170,000 MW molecule took 20

minutes, while the 22,000 MW material took 60 minutes (44).

Adsorption equilibria capacities also increased with decreasing

molecular size; the 170,000 MW material had an ultimate capacity

of 0.06 g/g, the 68,000 MW material's capacity being 0.11 g/g ,

and the 22,000 MW material's capacity being 0.19 g/g.

Adsorption equilibria were evaluated for different MW

fractions of dextran (a carbohydrate), with adsorption capacity

at a maximum for MW of 6000, higher MW fractions (up to 500,000)

resulted in lower loadings (45). Loadings of 400 mg/g were

obtained at equilibria concentrations of 10,000 mg/L.

To summarize, many small compounds (MW < 500) and very large

compounds (MW > 20,000) are not very adsorbable, whether they are

humic materials or fractionated wastewater components. The most

adsorbable fraction is the 1000 to 10,000 MW materials, with up

to 90% of it adsorbable. For many activated carbons, this size

material corresponds to the size material that adsorbs to a

maximum capacity when adsorbing a homologous series of compounds.

The adsorption rate in batch studies decreases with increasing

MW, with little or no removal expected in a "typical- filter-

adsorber used for water treatment (short contact time); although

in column studies all fractions are removed to some degree, even









if they were not removed in batch studies. Solution chemistry is

very important for compound adsorption, with calcium and

magnesium cations apparently the most important species.







EXPERIMENTAL

Materials: The following chemicals were obtained from

either Sigma Chemical, Aldrich Chemical, or Fisher Scientific for

use as received:

polysaccharides: sucrose (342 MW)
inulin (5200 MW)
xylan (20,000 MW)

protein: egg albumin (45,000 MW)

synthetics: phenol (94 MW)
polyethylene glycol (400 MW)
polyacrylic acid (5,000 MW)
polyvinyl pyrrolidone (10,000 MW)


The following activated carbons were used as provided by the

manufacturers:

Filtrasorb 400 Calgon Corporation

Hydrodarco 4000 ICI America

BAC-SQ Kureha Chemical Co. (Japan)

Filtrasorb 400 (F400) is a bituminous coal-based activated carbon

with an approximate surface area of 1100 m2/g; Hydrodarco 4000

(HD4000) is a lignite-based carbon with a surface area of 700

m2/g; and BAC-SQ is a petroleum pitch-based carbon with a surface

area of 1100 m2/g. The petroleum carbon has the smallest pores,

with 67 percent of its pore volume in pores smaller than 2

nanometers (Table I). F400 has 47 percent of its pore volume in

pores smaller than 2 nanometers, and HD4000 has only 2 percent of

its volume in these small pores, with 60 percent of its volume in

pores larger than 10 nanometers. If pore volume is the main

factor for the adsorptive capacity of the molecules and surface

chemistry is a minor consideration, the largest molecules should

be able to penetrate and adsorb better in the HD4000, then F400,









which would be better than BAC-SQ; conversely, the smallest

molecules should adsorb better in the BAC-SQ, followed by the

F400, followed by HD4000. All activated carbon for the

equilibria studies was pulverized in a Waring blender, sieved

through a U.S. Standard Sieve No. 120 (<125 microns), washed with

distilled water until the supernatant was clear, dried at 105 C

for 24 hours, and stored in sealed bottles until used.


TABLE I

Pore Volume Distribution for Adsorbents



Pore Size Percent of Total
nanometers Pore Volume

BAC-SQ F400 HD4000


< 2 67 47 2
2 3 11 22 12
3 5 7 10 11
5 10 6 8 15
> 10 9 13 60


MilliQ deionized water was used to prepare all solutions;

for the equilibrium studies using low initial solute concentra-

tions (10 20 mg/L), the deionized water was filtered through an

18-inch deep granular activated carbon bed prior to use, to

remove any organic compounds that may be present. No buffer was

used since some studies indicate their effects were small (46).

Equilibria Studies: Initial equilibria studies were

performed using an oscillating platform shaker operating at 150

rpm. Although the equilibrium time tests were based on this

system, the mixing pattern of the activated carbon in solution







did not appear satisfactory, as the activated carbon partially

settled in the center of the flask. Wooden racks were

constructed and mounted on a metal frame on the oscillating

shaker table, the racks held 120 ml Wheaton bottles at a 450

angle this was done to induce turbulence (Figure 1); also, a

drum was modified to hold wooden racks of Wheaton bottles, with

the drum then placed on a dual roller device to rotate at low

rpms (about 15 rpm was the slowest speed that our system could

maintain Figure 2). Both systems provided better mixing

patterns, with no difference noted among the isotherms performed

on all three systems.

Albumin was used in an initial test to determine the time

required for the largest compound (45,000 MW) to reach equilibrium.

During the first two hours of shaking, the albumin concentration

decreased from 60 mg/L to nearly 30 mg/L, and after 5 days it had

dropped to 12.5 mg/L. In the next 9 days, the concentration

decreased to 11 mg/L, so it appeared 14 days would be sufficient;

but to ensure sufficient equilibration, 21 days of mixing

were provided. This time is more than adequate based on a

comparison to the times used by others (one to seven days) who

studied large compound adsorption (23, 25, 28, 30, 36, 38).

At the completion of 21 days of mixing, the carbon was

settled and then a sufficient quantity of solution was filtered

through a Whatman GF/C filter pad; the albumin solution did not

filter well and required centrifuging rather than filtration to

remove the carbon fines from solution.

Column Studies: Fixed-bed adsorber studies were

performed in 25mm internal diameter glass columns of varying



























TABLE FOR ISOT~HE? ;R


F- 2 '7.T- 'TUS ,:. reTr Lr


.'.PE I iO.C-iFIED


SHAKER







lengths to provide the desired contact times. Filtrasorb 400 was

used and has an average particle diameter of 1 mm, resulting in

a 25:1 column to particle diameter ratio -- this is considered

by many to be sufficient to eliminate wall effects (47).

Masterflex tubing pumps were used to control the flow to the

adsorbers, with desired flows of 0.5, 1.0, 2.0, and 4.0 gpm/ft2.

Samples were taken after each column and analyzed for the

specific components as indicated below.

Analytical Methods: All single component equilibrium

samples were analyzed using an Oceanographics International Total

Organic Carbon Analyzer Model 525 B. Both the direct inject and

the ampule techniques were used; when the final equilibrium

solution concentration was greater than 5 mg/L, the direct inject

technique was used, when smaller values resulted, the ampule

technique was used. Standard curves were run each time analyses

were performed, with the standard prepared using the compound being

studied; so results are reported as mg/L as compound and not TOC.

When mixtures were used, each mixture contained a polysaccharide

(inulin or xylan) which could be determined directly using the

phenol-sulfuric acid test (50). Phenol was a second component

and ultraviolet spectroscopy (Perkin-Elmer dual beam spectro-

photometer Model 5200) was used to determine its concentration.

The third component in the mixture was calculated by measuring

the solution's TOC and subtracting the TOC response of the

equivalent concentrations of the polysaccharide and phenol and

then converting the remaining TOC response into the third

compound's concentration.








RESULTS AND DISCUSSION

Equilibria Tests:

Equilibrium isotherms were performed on the three carbo-

hydrates using all three activated carbons: F400, HD4000, BAC-SQ.

Since F400 is the most frequently studied adsorbent, it was used

to adsorb additional large compounds for comparative purposes.

The Freundlich adsorption equilibrium model was used to represent

the data, since it has been found to fit many data successfully.

Unfortunately it is an empirical model; but, a theory has been

proposed that results in the Freundlich isotherm equation (49).

The theory is based on adsorbing substances on a surface with

heterogeneous surface energies, therefore the heat of adsorption

varies with surface coverage. The Freundlich equation is:


1/n
X = K C
X=KC



Where, X is the loading of the solute on the adsorbent in mg/g;

C is the equilibrium solution concentration of solute in mg/L;

and K and n are empirical constants. K is proportional to the

adsorption energy, with the larger the value, the greater the

adsorption energy; and n is related to the intensity of

adsorption, with the greater the n, the more intensely the

material is adsorbed (49). The Freundlich isotherm is used to

represent the adsorption of the large compounds on the various

carbons, although its ability to fit the data was not always

good. The Freundlich constants are provided in Table II.







TABLE

FREUNDLICH


II

CONSTANTS


Compound


Sucrose



Inulin



Xylan


MW Adsorbent


342 F400
HD4000
BAC-SQ

5200 F400
HD4000
BAC-SQ

20,000 F400
HD4000
BAC-SQ


# of data


11
9
10

4
7
7

13
5
5


Freundlich
K, mg/g


.95
3.03
7.50

37.2
69.6
84.7

1.66
2.07
1.5E-06


Constants
n


.64
2.27
1.66

4.80
5.63
4.35

2.82
1.38
.29


PEG 400 400


PAA

PVP

Albumin


5000

10,000


F400
HD4000
BAC-SQ

F400

F400


40.8
29.1
18.8

6.1E-05

.01


3.89
4.38
1.53

.34

.48


45,000 F400 6 4.5E-04 .35

Phenol 94 F400 78.1 4.72
(from reference 46)

PEG polyethylene glycol
PAA polyacrylic acid
PVP polyvinyl pyrrolidone

A wide variation in both Freundlich K and n constants exists

for the large compounds tested. The adsorption energy term, K,

varies by nearly 8 orders of magnitude, with three compounds

being essentially non-adsorbable or poorly adsorbed on a

particular adsorbent: xylan was non-adsorbable on BAC-SQ; PAA,

PVP, and albumin were poorly adsorbed on F400. Inulin and PEG

400 were both adsorbed strongly on all carbons. Adsorption

intensity terms varied from .29 to 5.6, with values greater than









1 required for favorable adsorption (49). All three of the

additional large compounds adsorbed on F400 were poorly

adsorbed, as indicated by the unfavorable n terms. The xylan was

also unfavorably adsorbed on BAC-SQ, it probably could not

penetrate its narrow pore structure. In addition, sucrose's n

value is less than 1 for F400, but this is misleading since a

two-sloped isotherm plot resulted, and these values are for the

steep portion of the curve; a horizontal curve results at higher

concentrations.

Sucrose adsorption follows a pattern expected for the

adsorption of small molecules based on previous experience

(50) and published data using carbons with similar pore

structures, such as a coconut-based (Columbia carbon) or another

petroleum carbon (Witco carbons)(51, 52); the small pore

diameter petroleum-based carbon has the greatest adsorptive

capacity, followed by the bituminous-based, and, with a

considerably lower capacity, the lignite-based carbon (Figure

3). Since the petroleum and bituminous coal carbons have

similar surface areas, the main difference is in the pore size

distributions (or possibly the surface chemistry which was not

evaluated in this study) -- the smaller pores apparently exert

stronger adsorption forces on the solute, permitting greater

adsorption. With the F400 carbon, a two-sloped isotherm results,

with the maximum capacity obtained at an equilibrium solution

concentration of about 80 mg/L and no additional adsorption

occurring with increased concentrations. Multisloped isotherms

have been observed by others (53, 54). Sucrose adsorption









































CONCENTRATION,


* BAC- SQ
X F400
OHD4000


300


SUCROSE


mV/L


ISOTHERM


100


cD
z

Q

-J


FIGURE 3:


SUCROSE








follows the adsorption pattern expected for smaller compound

adsorption.

BAC-SQ's adsorptive capacity is nearly double F400's

capacity for inulin (MW 5200), a result that is similar to those

obtained for sucrose (Fiqure 4). The lignite carbon's capacity is

is considerably improved relative to the other carbons, since its

capacity for the inulin is greater than F400's and is about two-

thirds of BAC-SQ's (it was about 20% of BAC-SQ's capacity for

sucrose), Inulin's larger size (5200 MW) was expected to limit

its access into most of BAC-SQ's pores, and to a large amount of

F400's; so the lignite carbon was expected to perform best due to

its better performance for the adsorption of humic substances of

this size (21). Favorable adsorption behavior is exhibited with

good adsorption capacities for all three activated carbons. The

small pores of BAC-SQ are large enough to permit access of the

inulin, since approximately 40% of the total pore volume would be

occupied (if inulin's adsorbed specific gravity is 1.5). HD4000's

pores are better suited to the adsorption of this polysaccharide;

the larger molecule fills the pores more completely, leading to

stronger adsorption energies (K constant increased 20 times) and

therefore to greater capacities.

Xylan (MW 20,000) does not adsorb on BAC--SQ, while F400

and HD4000 capacities are considerably reduced relative to those

with inulin (Figure 5). The large-pore lignite carbon has a

much greater capacity than F400. This pattern was expected from

the pore size distributions and the results with humic

substances (21). With the very large molecules, to obtain better

adsorption, a larger pore size distribution is required to obtain












300-



200-






100o-
C, 80-

O
60
50-
-i
40-

BAC- SQ
30 X F400

0 H 4000

0 8 10 20 30 40 60 80 100 200


INULIN, m6/L


FIGURE 4: INULIN ISOTHERMS























X :X


X F400


0 HD4000


2 3 4 6 8 10 20


XYLAN, mq/L


40 60 80 100


FIGURE 5: XYLAN ISOTHERMS







the best adsorption capacity. But, its capacity is not

large and there may be better means of removing these compounds,

such as chemical coagulation or precipitation.

The small, narrow pore size distribution of BAC-SQ was

expected to provide excellent adsorptive capacity for the sucrose

molecule, with progressively poorer adsorption of the other

polysaccharides. Sucrose adsorption is favorable on this carbon,

with a maximum observed capacity of nearly 100 mg/g; but the 5200

MW inulin's adsorption is surprisingly good with a maximum

capacity of 300 mg/g and a nearly horizontal slope in a

Freundlich plot (Figure 6), indicating favorable adsorption.

The xylan's adsorption (MW 20,000) is basically non-existent,

which is not a surprise since it should be excluded from

this carbon due to the narrow range of pore sizes. Use of this

adsorbent finds the molecules of 5200 MW capable of penetrating

to pores smaller than 3 nanometers and adsorbing there, with this

adsorption stronger than for the smaller molecules; therefore,

this size material should be effectively removed from water and

wastewater in studies with all activated carbons. Studies with

fractionated materials found this fraction to be adsorbed best

(19, 37, 38).

Adsorption of the three polysaccharides on F400 exhibits the

same pattern, except the xylan adsorbs to an appreciable extent

(Figure 7). Inulin adsorbs best, with a maximum observed

capacity of 105 mg/g and a very shallow slope. Sucrose

adsorption follows the Freundlich pattern up to about 80 mg/L

equilibrium solution concentration, where its capacity plateaus

at 60 70 mg/g. The xylan's adsorption pattern is favorable,









300'


200'





100.
go-
80

60.



40


INULIN K


x


oQ o


SUCROSE


g 20-


CD
2
00

o O0 XYLAN
--
8-

6-



4-


3-


2
10 20 30 40 60 80 10


CONCENTRATION, gi/L



FIGURE 6: ISOTHERMS ON BAC-SQ




















40-


30.



0D 20-
Z


0
-J


X INULIN

o
X 0o

0SUROS

SUCROSE


S


XYLAN


6 8


CONCENTRATION, rng/L


FIGURE 7- ISOTHERMS ON F400








with a low adsorptive capacity (maximum loading about 20 mg/g).

F400 has more surface area and pore volume in the pores greater

than 2 nanometers than BAC-SQ (53% versus 33%), permitting the

larger molecule to penetrate somewhat into the adsorbent.

The adsorption pattern for the polysaccharides on HD4000 is

slightly different than for other adsorbents (Figure 8). Inulin

adsorption is again the best adsorbed compound, with a maximum

observed capacity of 180 mg/g. But, xylan adsorbs better

than sucrose on HD4000. This results from a combination of

the relatively good adsorption of xylan (maximum loading of 36

mg/g) and the poor loading of the sucrose (maximum loading of 28

mg/g). The large pore structure of the HD4000 does not provide

sufficient pore volume in the micropores for strong adsorption of

small molecules, while providing larger pores for stronger

xylan adsorption. Unlike the study with polyethylene glycols

(44), a maximum capacity was obtained with this lignite carbon,

and adsorbing larger similar molecules resulted in a lower

loading; therefore, high loadings are not necessarily expected

from the large-pored carbons.

The adsorption of additional large compounds was evaluated

on F400: PEG polyethylene glycol 400 (MW 400), PAA -

polyacrylic acid (MW 5000), PVP polyvinyl pyrollidone (MW

10,000), and albumin (MW 45,000) (Figure 9). As with the xylan

on BAC-SQ, the albumin was poorly adsorbed, with a vertical

isotherm near the starting solution concentration; its large size

prevented any appreciable adsorption. The PEG is favorably

adsorbed, with a Freundlich n of 3.89; adsorptive capacity at 100

mg/L equilibrium concentration was about 130 mg/g or considerably
























or


2
Z


0
-J


x
X

INULIN












d
XYLAN


SUCROSE


CONCENTRATION, Wm/L


FIGURE 8: ISOTHERMS


ON HD4000














0 PEG


PAA


10 20 30 40


CONCENTRATION,


200


MG/L


FIGURE 9: MISCELLANEOUS -COMPOUNDS


ON F400







greater than for sucrose's capacity of 66 mg/g at a similar

concentration for the similarly sized compounds. So other

factors in addition to size are important to its adsorption, as

is the case for smaller compounds frequently studied. Adsorption

of the 5000 MW PAA is unfavorable, with an n of .34; surface

charge groups on the PAA must be an important consideration since

this compound is not adsorbed in a manner similar to the

comparably sized inulin. The PVP adsorption pattern is also

unfavorable (n = .52), although the large compound's adsorption

is reasonably good (64 mg/g) at an equilibrium concentration of

60 mg/L. The larger PVP adsorbs better than the PAA, with its

capacity somewhat lower than the smaller inulin molecule, but

considerably greater than the larger xylan molecule. The upper

limit of the capacity obtained for the various large compounds

tested except for the PAA, indicates that the optimum size for

utilization of the F400 pore volume is about 5000, with either

decreasing or increasing compound size resulting in lower

adsorptive capacities. Note: a limited number of compounds from

different families were studied, and very large concentrations

were not tested to determine the actual limiting adsorptive

capacities. A more extensive study of large compound adsorption

is needed.

Column Tests:

The column studies were intended to monitor the breakthrough

pattern of three different molecular weight materials to

determine if size fractionation occurs within an adsorber. If

molecular weight size fractionation occurred, it was to be

determined if the smaller compound is able to displace the larger









compounds and result in chromatographic effects, similar to those

observed when small compounds are competing for adsorption

sites (55). The three compounds included inulin, a polysaccharide

that can be measured by the phenol-sulfuric acid test (48),

phenol, which could be measured by ultraviolet spectroscopy,

and polyethylene glycol (either 1000 or 14,000 molecular weight),

which would be determined by total carbon -- the total organic

carbon response of the equivalent amount of phenol and inulin

would be subtracted from the sample's TOC, with the remaining TOC

converted to PEG. Each compound was present at 75 mg/L. Four

flows (0.5, 1, 2, & 4 gpm/ft2) were to be used to determine the

superficial velocity and/or contact time effect on the breakthrough

pattern.

While the program appeared reasonable and should have

given insight into large compound adsorption, several problems

developed. First, the principal investigator left the

University of Florida in the middle of the project and the

remaining funds could not be transferred to his new university.

Second, the graduate student working on the equilibrium portion

was replaced for the column studies, and a student was employed

who was not dependent upon the study results for his graduation;

the principal investigator therefore had little control over the

remainder of the project. Third, several problems were

encountered with the total organic carbon analyzer and the backup

total organic carbon analyzer (Beckman 914), and the funds were

not available to repair either system. Fourth, when the

difficulty with the TOC analysis persisted, the samples were







not shipped, as requested, to the principal investigator's new

university so the TOC analyses could be performed (too much time

may have transpired anyway). Therefore, the results presented

here do not include three component interactions within the column,

since only two were analyzed.



The breakthrough pattern for phenol and inulin at 1 gpm/ft2

flow with PEG 1000 indicates that inulin breaks through before

phenol for the first and second columns (empty bed contact times

of 7.5, and 15 minutes) (Figure 10); this could be predicted from

Freundlich K values, since phenol's K is 78 and inulin's is 37

mg/g (55). But, when the phenol reached the third column, little

capacity remained for it, while the inulin continued to be

removed (contact time may have a significant effect on adsorption

results, but more data are needed). Around day eight when the

inulin broke through, a very steep breakthrough pattern was

obtained, with its effluent concentration exceeding phenol's

within two days. After 16 days, the fourth bed was still

removing nearly all the phenol, with a relatively constant

leakage of 3 to 4 mg/L of inulin. The PEG 1000 was expected to

adsorb well, with its capacity likely to be greater than

phenol's. How it breaks through the columns is not known, and

its effect on the phenol's reversal of order in column three is

not known; premature exhaustion may be occurring, but similar

results would have been expected in the fourth column.

The 1 gpm/ft2 run using the PEG 14,000 results in inulin

breaking through prior to the inulin in all three columns (Figure

11); the fourth column had to be removed from operation early in


























0 2 4 6 8 10 12 14 16 18 20


TIME


FIGURE 10- BREAKTHROUGH


CURVES


DAYS
2
FOR I GPM/FT,


PEG 1000 SYSTEM


0 2 4 6 8 10 12 14 16 18 20


TIME


DAYS


FIGURE II: BREAKTHROUGH CURVES


FOR I GPM/FT2


SYSTEM







the run due to a rapid headloss buildup. The carbon's capacity

for PEG 14,000 is expected to be less than for the phenol and

inulin due to its larger size; whether it did remains to be

determined. Times to breakthrough for both phenol and inulin are

given for the two systems (Table III). PEG size had no effect on

inulin breakthrough, even though the one PEG was smaller (MW =

1000, inulin = 5200) and the other was larger (MW = 14,000).


TABLE III

Times to Breakpoint
1 gpm/ft2 System


Compound



Inulin


phenol


Breakpoint
mg/L


10


Contact
Time, min


7.5
15
30


7.5
15
30

7.5
15
30

7.5
15
30

7.5
15
30

7.5
15
30


Time, days
PEG 1000 PEG 14,000


.7 .5
2.8 2.2
7.9 8.3


1.0
3.4
8.3

1.2
3.7
9.2

2.2
4.9
5.2

2.6
5.9
7.3

2.9
6.3
8.3


.7
3.5
8.9

1.0
3.8
9.5

2.1
5.8
12.2

2.6
6.3
14.7

2.8
6.7
15.7


This indicates that both PEGs are better or more poorly adsorbed

than the inulin, since the inulin apparently is seeing the same

number of adsorption sites within the columns. With phenol, a


__~~I~








difference is noted in the breakthrough pattern of the phenol.

At the greater contact times, the phenol competes better for

adsorption sites in the system with the higher MW PEG; this

indicates that the larger PEG is not adsorbing as well as the

smaller one at deeper locations in the adsorber. When the phenol

reaches these points the adsorbent still has plenty of adsorption

sites left for the phenol. Or conversely, all the larger PEG

could be adsorbing on the same carbon already loaded with phenol

and then the phenol doesn't have to compete for adsorption sites

with the smaller PEG for sites; the first hypothesis is more

likely due to kinetic considerations.

The same two chemical systems were evaluated at 2.0 gpm/ft2

(Figures 11 & 12). Again, the inulin breaks through the adsorbers

before the phenol, in fact it has broken through the 7.5 minute

empty bed contact time column before the phenol has broken

through the 3.75 minute columns. Breakthrough for both the

phenol and the inulin is rapid, with less than a day needed from

the time breakthrough begins. The times to different breakpoints

were determined for the different PEGs (Table IV).

The different PEGs have no apparent effect on either phenol

or inulin breakthrough at this higher superficial velocity, and

longer contact times were not tested.

At 4 gpm/ft2, the headless was too great for our system and

within three days the runs were abandoned with the 7.5 minute

contact time systems performing the same for both PEGs. The 0.5

gpm/ft2 runs were not performed, and therefore the longer contact

times were not tested. The greatest effects of flow rate were

expected to be observed at the slowest flow and greatest contact







70


60


50.
0

S40

Z
LW 30
0
0
0 20


0 I 2 3 4 5 6 7 8 9 10
TIME. DAYS


FIGURE 12: BREAKTHROUGH CURVES FOR 2 GPM/FT2


70-


-J
N 60-


Z50-
0

< 40-
CE

Z 30
0

020
0)


"/3,75


SYSTEM, PEG 1000


/I


15 MIN
CONTACT
TIME N/
/


/


- INULIN
--- PHENOL


PEG 14,000
F400


I I I I I I I I I I I
0 I 2 3 4 5 6 7 8 9 10


TIME, .DAYS


FIGURE 13: BREAKTHROUGH CURVES FOR 2 GPM/FT2


I
I
I
I
I


o
75/
//

I
/
I
I
I


I5-MIN. /
CONTACT -- INULIN
TIME I ---PHENOL
/

I
/
I
I
I
I

I PEG 1000
//- F400


SYSTEM, PEG 14000








time, since it was under similar conditions that the premature

breakthrough was observed (40).


TABLE IV

Times to Breakpoint
2 gpm/ft2 system

Breakpoint Contact B
mg/L Time, min


10 15

20 15

30 7.5
15

10 3.75
7.5
15

20 3.75
7.5
15

30 3.75
7.5
15


breakthrough Time, days
PEG 1000 PEG 14,000


3.4

3.8

1.1
4.1

1.1
2.6
6.6

1.3
3.1
6.8

1.5
3.3
7.1


3.1

3.3

.6
3.4

.9
3.0
6.4

.9
2.9
6.8

1.6
3.6
7.1


Activated carbon use rates were calculated to the 30 mg/L

breakpoint of both phenol and inulin for the several contact

times at each flow (Figure 14). With inulin, little difference

in carbon usage is observed between the PEG 1000 (solid line) and

the 14,000 (dashed line) at all contact times when the flow is 1

gpm/ft2; but, when the flow is 2 gpm/ft2, a substantial

difference was observed at the 7.5 minute contact time. The

larger PEG caused the inulin to breakthrough much faster,

increasing the carbon usage; whether the inulin moved with the

PEG or the PEG blocked the adsorption sites is unknown. Phenol


Compound



Inulin






phenol

















- PEG 1000
---PEG 14,000


30-




25-


0


a -
0 -








0

5-
ff


PHENOL


-~~ ~ -O -r -


2
0 2 GPM / FT
I GPM / FT

0 5 10 1S 20 t5 30

CONTACT TIME, MINUTES


FIGURE 14: CARBON USAGE TO A 30MG/L BREAKPOINT


INULIN

% \
%.'








breakthrough is unaffected by the PEG at the 2 gpm/ft2, but at

the 30 minute contact time with the 1 gpm/ft2, premature

breakthrough is possibly observed with the smaller PEG. The

smaller PEG probably proceeded deeper into the adsorber and

blocked the pores as hypothesized in the premature breakthrough

article (19). More contact time studies need to be performed

as well as determination of the PEG breakthrough.

Contact time has a significant effect on the carbon usage

for the larger inulin molecule, with contact times less than 15

minutes causing substantially greater carbon usages (Figure 14).

The carbon usage for phenol breakthrough is not affected

considerably by contact time, except in the case for the smaller

PEG at the long contact time. So in the case of one molecule,

longer contact times are needed for better removal, while for the

smaller molecule this may prove detrimental.



CONCLUSIONS

1. The 5200 MW polysaccharide (inulin) adsorbed much better

on all three adsorbents than did the 342 MW sucrose or

the 20.000 MW xylan, as is evidenced by both the Freundlich

isotherm adsorption energy term ( K constant) and the

adsorption intensity term (n).

2. A small-pored acitvated carbon (BAC-SQ) has little

adsorption space for large MW compounds (20,000 MW).

3. Even for the large-pore activated carbon (HD4000), pore

volume accessibility is limited, since xylan (20,000 MW)

capacity is much less than for inulin (5200 MW), unlike

when adsorbing a series of PEGs.








4. No chromatographic effect was noted for phenol or inulin

in column runs at 1 or 2 gpm/ft2 with either PEG.

5. Inulin breaks through the column before phenol at both

flows, as would be expected from the Freundlich K values.

6. Carbon use decreases with increased contact time for

inulin in both flow systems.

7. Carbon use decreases with increased contact time for

phenol at 2 gpm/ft2 for both PEG systems.

8. Premature exhaustion may occur at longer contact times

(30 minutes) at 1 gpm/ft2, since carbon use increased

for the PEG 1000 system.








REFERENCES


0. Nelson, A.J. "Removal of Large-Molecular Weight Organic
Compounds by Activated Carbon from Dilute Aqueous Solution",
University of Florida, Department of Environmental
Engineering Sciences, Dec. 1982

1. Committee Report, V.L. Snoeyink Chairman, "Measurement and
Control of Organic Contaminants by Utilities" JQ uJrn. Am.
ater Works Assn., _69, 5, 267-271, 1977

2. Hager, D.G. "Activated Carbon Used for Large Scale Water
Treatment" Environ. Sci. & Technol., 1, 4, 287-291, 1967

3. Wood, P.R.; DeMarco, J.; "Chapter 5: Removing Total Organic
Carbon and Trihalomethane Precusor Substances" ActivaAe
Carbon Adsorption of Organics from the ~AeoasE ha, Vol.
2, (M.J. McGuire and I.H. Suffet editors), Ann Arbor Science,
Ann Arbor, MI., pp. 115 135, 1980

4. Brodtmann, N.V., Jr.; DeMarco, J.; Greenberg, D. "Chapter 8:
Critical Study of Large-Scale Granular Activated Carbon
Filter Units for the Removal of Organic Substances from
Drinking Water" Ibid., pp. 179 222

5. Yohe, T.C.; Suffet, I.H.; Cairo, P.R. "Specific Organic
Removal by Granular Activated Carbon Pilot Contactors" Journ.
Am. Water Works Assn., 73, 8, 402-410, 1981

6. Environ. Protection Agency, "Control of Organic Contaminants
in Drinking Water" Federal Register, 43, 28, 5756-5758, 1978

7. Pendygraft, G.W.; et al "The EPA-Proposed Granular Activated
Carbon Treatment Requirement: Panacea or Pandora's Box?"
Journ. Am. Water Works Assn., 71, 2, 52-60, 1979

8. O'Brein, R.P.; Jordan, D.M.; Musser, W.R. "Trace Organic
Removal from Contaminated Groundwaters with Granular
Activated Carbon" Paper, Am. Chem. Soc. Mtg., 3/29 4/3/81,
Atlanta, GA

9. Anon, "The Calgon Adsorption Service" Calgon Corp. Brochure

10. Slechta, A.F.; Culp, G.L. "Water Reclamation Studies at the
South Lake Tahoe Public Utility District" Journ. Water
Pollution Control Fed., 39, 5,787-814, 1967

11. McCarty, P.L.; et al "Water Factory 21: Reclaimed Water,
Volatile Organics, Virus, and Treatment Performance" EPA-
600/2-78-076, Environ. Protect. Agency, Washington, D.C.,
June 1978

12. Weber, W.J., Jr.; Hopkins, C.B.; Bloom, "Physicochemical
Treatment of Wastewater" Journ, Water Pollution Control
F_d., 42, 1, 83-99, 1970








13. Hager, D.G. "A Survey of Industrial Wastewater Treatment by
Granular Activated Carbon Part II", Paper at 68th Annual Am.
Instit. of Chem. Engrs. Mtg., Nov. 1975, Los Angeles, CA

14. Environ. Protection Agency, "Preliminary Assessment of
Suspected Carcinogens in Drinking Water" Report to U.S.
Congress Office of Toxic Substances, Washington, D.C., Dec. 1975

15. National Academy of Sciences, "Drinking Water and Health", Safe
Drinking Water Committee, National Research Council,
Washington, D.C., 1977

16. Black, A.P.; Christman, R.F. "Chemical Characteristics of
Fulvic Acids" Journ. Am. Water Works Assn., 55, 7, 897-912,
1963

17. Gjessing, E.; Lee, G.F. "Fractionation of Organic Matter in
Natural Waters on Sephadex Columns" Environ. Sci. & Technol.,
1, 8, 631-638, 1967

18. Schoor, J.L.; et al "Trihalomethane Yields as a Function of
Precursor Molecular Weight" Environ. Sci. & Technol., 13, 9,
1134-1138, 1979

19. Arbuckle, W.B. "Chapter 10: Premature Exhaustion of Activated
Carbon Columns" Activated Carbon Adsorption of Organics from
the Aqueous Phase, Vol. 2, (M.J. McGuire and I.H. Suffet
editors), Ann Arbor Science, Ann Arbor, MI.,pp.237-252,1980

20. Davis, J.A.; Gloor, R. "Adsorption of Dissolved Organics in
Lake Water by Aluminum Oxide. Effect of Molecular Weight"
Environ. Sci. & Technol., 15, 10, 1223-1229, 1981

21. Lee, M.C.; Snoeyink, V.L.; Crittenden, J.C. "Activated
Carbon Adsorption of Humic Substances" J. Am. Water Works
Assn., 73, 8, 440-446, 1981

22. Manos, G.P.; Tsai, C-E "Mechanisms of Humic Material
Adsorption on Kaolin Clay and Activated Carbon" Water, Air,
Soil Pollution, 14, 419-427, 1980

23. McCreary, J.J.; Snoeyink, V.L. "Characterization and
Activated Carbon Adsorption of Several Humic Substances"
Water Research, 14, 2, 151-160, 1980

24. Boening, P.H.; Beckmann, D.D.; Snoeyink, V.L. "Acitvated
Carbon versus Resin Adsorption of Humic Substances" Journ.
Am. Water Works Assn., 72, 1, 54-59, 1980

25. Weber, W.J., Jr.; et al "Chapter 16: Potential Mechanisms
for Removal of Humic Acids from Water by Activated Carbon"
Activated Carbon Adsorption of Organics from the Aqueous
Phase, Vol. 1 (I.H. Suffect & M.J. McGuire editors), Ann
Arbor Science, Ann Arbor, MI., 317-336, 1980









26. Randtke, S.J.; Jepsen, C.P. "Effects of Salts on Activated
Carbon Adsorption of Fulvic Acids" Journ. Am. Water Works
Assn., 74, 2, 84-93, 1982

27. Randtke, S.J.; Jepsen, C.P. "Chemical Pretreatment for
Activated-Carbon Adsorption" J. Am Water Works Assn, 73, 8,
411-419, 1981

28. Youssefi, M.; Faust, S.D. "Chapter 5: Adsorption and
Formation of Light Halogenated Hydrocarbons and Humic Acid
in Water by Granular Activated Carbon" Activated Carbon
Adsorption of Organics from the Aqueous Phase, Vol. 1, (I.,H.
Suffet & M.J. McGuire editors), Ann Arbor Science, Ann Arbor,
MI., 133-143, 1980

29. Murin, C.J.; Snoeyink, V.L. "Competitive Adsorption of 2,4-
dichlorophenol and 2,4,6-trichlorophenol in the Nanomolar to
Micromolar Concentration Range" Environ. Sci. & Technol., 13,
3, 305-311, 1979

30. Huang, J-C; Garrett, J.T., Jr. "Effects of Colloidal
Materials and Polyelectrolytes on Carbon Adsorption in
Aqueous Solution" Proc. Purdue Industrial Waste Conf., 30th, Ann
Arbor Science, Ann Arbor, MI., 1111-1121, 1976

31. Ward, T.M.; Getzen, F.W. "Influence of pH on the Adsorption
of Aromatic Acids on Activated Carbon", Environ. Sci. &
Technol., 4, 1, 64-67, 1970

32. Lee, M.C.; et al "Design of Carbon Beds to Remove Humic
Substances" J. Environ. Engr. Div., Am. Soc. of Civil
Engrs., 109, EE3, 631-645, 1983

33. Zuckerman, M.M.; Molof, A.H. "High Quality Reuse Water by
Chemical-Physical Wastewater Treatment" Journ. Water
Pollution Control Fed., 42, 3, 437-456, 1970

34. Guiruis, W.; et al "Improved Performance of Activated Carbon
by Pre-ozonation" J. WatIr Pollution Control Fed., 50, 2,
308-320, 1978

35. Gerba, C.P.; et al "Adsorption of Poliovirus onto Activated
Carbon in Wastewater" Environ. Sci. & Technol., 9, 8, 727-
731, 1975

36. Summers, R.S.; Roberts, P.V. "Simulation of DOC Removal in
Activated Carbon Beds" paper submitted to J. Environ. Engr.
Div., Am Soc. Civil Engrs., 1983

37. Kim, B.R.; Snoeyink, V.L.; Saunders, F.M. "Influence of
Activated Sludge Cell Retention Time on Adsorption" J.
Environ. Engr. Div., Am Soc. Civil Engrs., 102, EEl, 55-70,
1976







38. DeWalle, F.B.; Chian, E.S.K. "Removal of Organic Matter by
Activated Carbon Columns" J. Environ. Engr. Div., Am. Soc.
Civil Engrs., 100, EE5, 1089-1104, 1974

39. Chow, D.K.; David, M.M. "Compounds Resistant to Carbon
Adsorption in Municipal Wastewater Treatment" J. Am. Water
Works Assn., 69, 10, 555-561, 1977

40. Sherm, M.; Lawson, C.T. "Activated Carbon Adsorption of
Organic Chemical Manufacturing Wastewaters After Extensive
Biological Treatment" Proceedings, The Third National
Conference on Complete Water Reuse, Am. Inst. Chem.
Engrs./Environ. Protection Agency, pp. 196 203, 1976

41. Volesky, B.; Roy, C. "The Effect of Polysaccharidic Gums on
Activated Carbon Treatment of Textile Wastewaters" Water
Research, 13, 8, 791-800, 1979

42. Wang, L.K.; et al "Adsorption of Dissolved Organics from
Industrial Effluents onto Activated Carbon" Journ. of Appl.
Chem. Biotechnol., 25, 491-502, 1975

43. Wu, R-Y.A.; Okos, M.R. "Adsorption of Proteins onto
Activated Carbon and Phenolic Resins" paper 1982 Annual Am.
Inst. Chem. Engrs. Meeting, November 1982, Los Angles, CA.

44. Suzuki, M.; Kawai, T.; Kawaqoe, K. "Adsorption of
Poly(oxyethylene) of Various Molecular Weights from Aqueous
Solutions on Activated Carbon" J. of Chem. Engr. of Japan,
9, 3, 203-208, 1976

44. Claesson, I.; Claesson, S. "The Adsorption of Some High
Molecular Substances on Activated Carbon" Arkiv for Kemi,
Mineral. och Geol., 19, 5, 1-12, 1944

45. Ingelman, B.; Halling, M.S. "Some Physico-chemical
Experiments on Fractions of Dextran" Arkiv for Kemi, 1, 10, 61-80, 1949.

46. Peel, R.G.; Benedek, A. "Attainment of Equilibrium in
Activated Carbon Isotherm Studies" Environ. Sci. & Technol.,
14, 1, 66-71, 1980

47. Gilliland, E.R.; Baddour, R.F. "The Rate of Ion Exchange"
Ind. & Engr. Chem., 45, 2, 330-337, 1953

48. Benefield, L.D.; Randall, C.W. "The Phenol-Sulfuric Acid
Test: Effective Alternative for Carbohydrate Analysis" Water
& Sewage Works, 2, 55, 1976

49. Adamson, A.W. "Chapter IX: The Solid-Liquid Interface -
Adsorption from Solution" Physical Chemisty of Surfaces,
3rd ed., Wiley-Interscience, New York, 386-394, 1976

50. Arbuckle, W.B. Unpublished reports Union Carbide Corp.








51. Zagorski, J.S.; Faust, S.D. "Removal of Phenols from
Polluted Waters" Water Resources Research Inst., Trc.i~j.
Pe-?:c r No. A-033NJ, June 1974

52. Mullins, R.L., Jr.; et al "The Effectiveness of Several
Brands of Granular Activated Carbon for the Re .m.al of
Trihalomethanes from Drinking Water" Activated Carbon
Adsorption of_ rganics from the Aguquqo Pha1s.e, LIoB l, (I.H
Suffet & M.J. McGuire editors), Ann Arbor Sci., Ann Arbor,
MI, pp.273-307, 1980

53. Snoeyink, V.L.; Weber, W.J., Jr.; Mark, H.B., Jr. "Sorption
of Phenol and Nitrophenol by Activated Carbon" Environ. Sci.
& Technol., 3, 10, 918-926, 1969

54. Al-Bahrani, K.S.; Martin, RoJ. "Adsorption Studies Uxc.:-'
Gas-Liquid Chromatography I. Effect of Molecular Structure"
Water Research, 10, 731-736, 1976

55. Arbuckle, W.B.; Romagnoli, R.J. "Prediction of the
Preferentially Adsorbed Compound in Bisolute Column Studies",
Chem. Engr. Proq. Symp. Ser. Water-1979, 76, 197, 61-71,
1980




Full Text

PAGE 1

WATER IiRESOURCES researc center Publication No. 74 ADSORPTION OF LARGE COMPOUNDS ON ACTIVATED CARBON By w. B. Arbuckle Department of Environmental Engineering Sciences University of Florida Gainesville UNIVERSITY OF FLORIDA

PAGE 2

ACKNOWLEDGEMENTS The research support of the Florida Water Resources Research Center under the direction of Dr. James P. Heaney with funds provided by the U. S. Department of the Interior through their annual cooperative program .(OWRT Project No. A-045-FLA) and the Engineerihg and Industrial Experiment Station of the University of Florida, Gainesville, FL. is greatly acknowledged. In addition, Amy Jo Nelson's laboratory work on the adsorption equilibria and her master's degree special project report (reference 0) are acknowledged, as is the laboratory work of Jon Earle on the column studies.

PAGE 3

TABLE OF CONTENTS ';;" ..,; Acknowledgements. ...................... ............... i List of Tables ....................................... List of Figures ..... .. . . . .. iii Abstract . '".e, ."' .;" iv Introduction . .. .... ..... .. .. ............... 1 B a c k 9 r 0 U n d '. '. .'. ,'.. -.' :. ." .". 4 Humic substances . ..................... 5 Wastewater materials. .... ............. ....... ,"":.: 7 Pure compounds .. 46 10 Experimental . . . . . . .............. 13 Materials . . . . . . ........... .. 13 Equilibria studies .. : 14 Column studies .............. 15 Analytical methods . ........................ 17 Results and Dfscussion. e.' '.' 18 Equilibria tests . ., .. ; ...................... 18 Column tests ................................ ..... 31 Conclusions. .......... ................... ......... 40 References . . . . .. . . . . . 42 ; ;

PAGE 4

LIST OF TABLES Table I Pore Volume Distribution for Adsorbents .... 14 Table II -Freundlich Constants ....... 19 Table III Times to Breakpoint 1 gpm/ft2 system .... 35 Table IV Times to Breakpoint -2 gpm/ft2 system ... 38 LIST DE' FIGURES Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 Figure 14 --Modified Shaker Table for Isotherms 16 Roller Apparatus for Isotherms .. 16 Sucrose .............. 21 Inulin Isotherms ............... 23 Xylan Isotherms .......... 24 Isotherms on BAC-SQ ...... 26 Isotherms on F400 ...... 27 Isotherms on HD4000 .. 29 Miscellaneous Compounds on F400 ...... 30 Breakthrough Curves for 1 gpm/ft2, PEG 1000 System.o ............ e 34 Breakthrough Curves for 1 gpm/ft2, PEG 14,000 eDe e e.ee. 34 Breakthrough Curves for 2 gpm/ft2, PEG 1000 System 0 0 e ............ e III e a OIl 37 Breakthrough Curves for 2 gpm/ft2, PEG 14,000 Sy stem III _8 III III e ell. CI Q 37 Carbon Usage to a 30 mg/L Breakpoint .. 39 iii

PAGE 5

ABSTRACT. As Florida's water demand increases due t.O population growth and to increased per capita consumption, less desiraole sources of water will be used, requiring additional treatment for the removal of organic compounds. In addition, wastewaters will require additional treatment to remove organic compounds. In both cases, activated carbon adsorption is the process. most likely to be applied. Most of the organics present. in natural waters and biologically treated wastewaters.are large compounds (i.e. molecular weight [MW] > 500), and little' is known about large compound adsorption. An optimum size of adsorbate appears to exist, since the 5200 MW inulin adsorbs better than the 342 MW sucrose and the 20 ,000 MW xylan on all three adsorbents eval uated, including the small-pored petroleum pitch carbQn. When adsorbing the xylan, even the large-pored lignite carbon's capacity was much reduced. No chromatographic effect was noted in continuous column studies at I and 2 gpm/ft2. Premature exhaustion may exist as proposed in previous studies, but additi6nal stUdies are needed at longer contact times (> 30 minutes) ; in. general, carbon use decreases with increased contact time.' More studies are needed. iv

PAGE 6

INTRODUCTION Water and wastewater problems are likely to become very important issues in Florida in the near future; due to its rapidly expanding population. The problem is intensified since most of the population growth is occurring in the southern coastal 'regions of the state where the least potable water is available. Groundwater is used extensively for water supply, but additional water withdrawals may threaten existing supplies by permitting salt water to intrude into the fresh water aquifers, rendering them useless. Due to the large number of hazardous waste disposal sites within the state, the potential for .contaminating the groundwater is great. And some water supply wells have been abandoned due to pollution from the surface. Groundwater supplies are limited, and diligence is needed to protect the existing systems. Many communities use .surface waters to supply their water; surface water sources could be used more extensively in Florida, but water quality problems are likely to result, as well as legal problems associated with transferring water from one drainage basin to another, since most of the surface water is in the northern part of the state. Many surface waters are highly colored due to decaying vegetation, the color is not only aesthetically objectic;mable, but produces trihalomethanes, which are believed to be carcinogenic compounds. Special treatment will be required to remove this color,' with both activated carbon adsorption and chemical coagulation the most likely processes. In addition, surface water use will require protection from wastewater discharges, especially as the population expands. I

PAGE 7

More extensive wastewater treatment will be required when the waste is discharged into a stream used for potable water supply or when the waste is used to recharge the groundwater, whether to protect it from salt water intrusion or as an ultimate disposal method. Undoubtedly, concern for trace organic comrounds in these wastes will require activated carbon adsorption as a treatment step. Activated carbon adsorption is a very effective process for removing organic compounds from water. In potable water treatment, activated carbon has been used for to remove taste and odor causing compounds and for removing color bodies (1). Whenused in these applications, powdered activated carbon is usually added in the normal water treatment processing steps; if the water has extremely high levels of contaminants, then granular activated carbon has been used to replace the sand in the rapid sand filters (2) --providing better, more economical treatment. Recently adsorption has been extensively evaluated for the removal oftrihalomethaneprecursors and synthetic organic chemicals rom our potable. waters (3,4,5), with the U. S. Environmental Agency proposing it as a treatment requirement for all water plants providing water to communities greater than 75,000 people (6). For these applications, an additional processing step would probably be added to the treatment system, rather than replacing the sand with activated carbon as is done for-high levels of taste and odor producing compounds and for high color levels, adding substantially to the capital and operating costs of the water treatment system (7). 2

PAGE 8

When a groundwater becomes contaminated by synthetic chemicals and the water source is not abandoned, the water can be processed to acceptable quality using granular activated carbon systems (8). Portable systems can be taken to the site and installed in a short time, providing excellant removal of many compounds (9). In wastewater treatment, activated carbon adsorption is used to remove refractory compounds when a high quality water is needed such as for South Lake Tahoe (10) or Water Factory 21-(11). It is al so used with chemical addition to repl ace "normal" secondary treatment plants with physical/chemical treatment processes (12), where the adsorption system is used to remove the organic compounds. Industry can also use activated carbon to remov toxic and nondegradable compounds before discharging them into sewers (13), so they won't destroy or inhibit the biological treatment system --industrial applications are likely to increase dramatically if the U.S.E.P.A. sets strick limits on toxic compound discharges. In many potable water supplies, a variety of organic compounds exist in the water, with over 250 chemicals in the nation's potable water supplies (14). Extensive efforts are underway to identify the components of a water, but only a small fraction of the total organic of a water is identifiable (15). These are small molecules (smaller than 200 atomic units). The remaining compounds are frequently classified as humic substances which include both humic and fulvic acids (humic acids are less soluble under acidic conditions). Their molecular weight distribution has been evaluated using gel chromatography (16, 17, 18). Most of the 3

PAGE 9

compounds have a molecular weight greater than 500, with many being considerably latger than 10,000. In wastewaters, only a small fraction of the organic carbon present is identifiable as specific compounds --most of the remaining compounds are large compounds. Attempts have been made to identify the families of compounds, with proteins, carbohydrates, lignins, and tannins sometimes identified along with the molecular weight distributions. The purpose of this study is to provide basic background adsorption equilbria data for some specific high molecular weight model compounds; this should indicate which families of compounds could potentially be a problem adsorption processes. In addition, how well the model compounds adsorb on different activated carbons will be evaluated to determine if a particular carbon appears better suited for large compound adsorption. Fixed-bed adsorption studies will also be performed with a mixture of compounds to determine if the compounds fractionate by molecular weight within the adsorber. [This could be due to the slower diffusion of the larger compounds, causing the large compounds to penetrate deeper into the bed where they adsorb and block the adsorbent's pores; thereby preventing additional adsorption when the smaller compounds reach that point in the column. This phenomenon is referred to as premature exhaustion and results in larger than expected carbon usage (19)J. BACKGROUND There is not much information available on the adsorption of large molecular weight compounds, most of the available 4

PAGE 10

information is for humic sUbstances or fractionated wastewater. In addition, some information is contradictory. Humic Substances: These materials are very complex, source-dependent organic substances. They are classified as humic or fulvic acids, with the humic acids being less soluble compounds are separated from the fulvics by lowering the pH to 2, where the precipitate that forms is humic acid. The large fraction (MW > 4000) of lake water was found to adsorb better on aluminum oxide solids than the small fractions (MW < 400); their adsorption is more pH dependent (20). Adsorbing fulvic acids extracted from Michigan peat onto activated carbon finds the large materials (MW > 50,000) adsorbed less than the smaller ones (MW < 1000). There are indications that the small fractions associate with the large ones in solution when adsorbing, since the urifractionated sample's adsorption pattern was similar to that of the large fraction's (21). The adsorption of the small compounds correlates with the pore volume in pores smaller than 7 nanometers for the nine activated carbons tested, and the large compounds's adsorptive capacity correlates with the pore volume in pores smaller than 40 nanometers. Carbons with most of their pore volumes in the larger diameter pores would be more suitable for adsorbing large compounds. .Both adsorption capacity and adsorption rate increased with decreasing molecular weight for the fulvics studied. Other researchers have also found adsorptive capacity to increase with decreasing molecular weight (22, 23). Unfractionated soil fulvic acid adsorbed similar to the smallest molecular weight fraction (MW < 5000), while the soil hUmic acid's adsorption was similar to the 5

PAGE 11

middle molecular weight fraction's (5,000 50,000) (23). Removals depend upon the source of humic materials -(24). The solution's chemical conditions significantly affect the results. Humic acid was 10 times'more adsorbablefrom tap water than from deionized'-distilled water, indicating that co-and/or counter ions are important factors (25). Magnesium ions enhanced adsorption capacity more than cal.cium; hypochlorite ions .also increased capacity. Others have found little adsorption of humics in the absence of sal ts, wi th the anions having no effects (26). The cations were most important, with Ca enhancing adsorption more than Mg which had a greater effect than Na. Cation effects were much reduced at low pH values. In another system Phosphate was found to greatly increase adsorption capacity (2 to 3 times) (23). The after precipitation'of humic materials with alum, adsorb substantially better than those present before treatment (27, 21). Solution pH is also important, with studies finding capacity to increase with detreasing pH from 11 to 2 (23) while others found an optimum pH when using an aluminum oxide adsorbent instead of activated carbon (20) Humics were found to have an optimum adsorption pH in a study testing only 3 pHs, since the lowest pH (4.5)' resulted in a lower adsorptive capacity than the two higher pHs (28). Humic substances have a significant adverse effect on the adsorption of chlorophenols (29), while only exhibiting a slight effect on the adsorption of phenol at neutral pH (30). A larger adverse effect on phenol adsorption resulted at pH 9, but this is 6

PAGE 12

probably due to the ionization of the phenol as its pKa is approached (31). At pH 2 ,no interference with adsorption was observed (30), the humic materials would be expected to adsorb to a greater extent at pH 2 since they become insoluble; but the humic materials were believed to become insoluble colloids and therefore did not adsorb in the normal sense, but behaved in a manner similar to other colloids (30). Other colloids (clay or polyelectrolytes) added to the phenol system did not interfere wi th phenol adsorption by adsorbing and blocking the pores (30). Few adsorption column studies were reported. In one study a short empty-bed contact time was used (1.6 minutes, approximately equal to a 5inch deepadsorber with a flow of 2 gpm/ft2) and a "rapid" humic acid breakthrough was observed with only a small amount of material adsorbed (0.23 mg/g) (28). Humic substance adsorption has been modeled, with only 40 percent of the carbon's equilibrium capacity utilized at 80% breakthrough due to slow adsorption (32). Model studies found the system insensitive to the film transfer coefficient and, in the early stages of the breakthrough curve, to the surface diffusion coefficient; during the later stages of breakthrough, the system becomes sensitive to surface diffusion. Pretreatment of the humic substances with' alum increases the service time of the adsorber by 14 to 22 times (32)! Because of the slow surface diffusion, long empty-bed contact times are required to contain the masi transfer zone. Wastewater Materials: Several studies exist using waste-water compounds. In an early study, compounds larger than 1200 MW were claimed not to adsorb, with only the compounds about 400 7

PAGE 13

MW and smaller adsorbed (33). They hydI'olyzed the lar:ger compounds to smaller ones using lime and a sufficient reaction time; then the hydrolyzed organics were adsorbed with far better treatment resulting. Similar results were obtained in another physical-chemical scheme eval uated at the pilot level (34) Q In addition, a study on virus removal found the organic removal to increase substantially after lime treatment, like in the hydrolysis treatment (35). Many others have found the smaller compounds to account for most of the material (36, 37,38, 39). These studies were more detailed research studies; they all used secondary eff I uents.The studies showing the small materials to be more adsorbable used physical/chemical pilot plant influents and effluents, except for the virus study. which used a secondary effluent. In column studies, all MW fractions were found to be able to some extent, with 52 percent of the smallest compounds and 35 percent of the larger compounds the inta ate size was 90 percent removed (38). The largecompol1nds (MW > 50(000) that were not adsorbed were carbohydrate-like materials. Column studies performed on size-fractionated 01 cally-treated wastewaters found the smallest compounds not to orb, but, if the material was fractionated after all fractions adsorbed to some extent in the col umns (37) 0 Mean ce 11 residence time greatly affects the quantity of the largest compounds, with decreasing mean cell residence times resulting in more large compoundsv and therefore poorer overall removals by adsorptiono 8

PAGE 14

Batch kinetic and equilibria studies were performed on three fractions of an extensively treated industrial wastewater (60 hours hydraulic residence time, sludge age greater than 10 days) (19).. This waste was used in a previous pilot study where adding a second column in series with the first adsorber resulted in a 70 % increase in carbon usage, rather than the expected reduction in use (40); it was referred to as premature exhaustion. Pilot scale ultrafiltration and reverse osmosis was used to fractionate the waste prior to the adsorption studies. A majority of the material (48%) passed through the ultrafiltration membrane, but was rejected by the reverse osmosis membrane (about 342 to 6500 -MW). This fraction adsorbd best (79%), while the smallest material adsorbed least (55%); the large fraction was 62% adsorbable (19). Adsorption energy increased with increasing molecular weight, indicating that the larger compounds are more strongly adsorbed and 'therefore are more difficult to displace. The larger compounds adsorb at slower rate, as would be expected. This study proposed that the large compounds penetrate far deeper into the adsorber than thesmaller compounds before adsorbing. Deep in the adsorber, there is less competition for adsorption sites and they slowly adsorb, but over a period of time they have utilized a large portion of the adsorbent's capacity. Due to their size, they effectively block the pores of the carbon and since they are stongly adsorbed, they cannot be displaced. Gums were tested in dye wastes for their effects on adsorption, with the large gums (MW > 200,000) having a positive effect on the activated carbon system at 65 to 100 mg/L levels, but having an adverse effect at lower concentrations --no 9

PAGE 15

reasons were given (41). The effect of pH on the adsorption of lignosulphate and tannins were eval.uated in real industrial wastes (42), with 1 ignin and tannins both removed be.tter from their respective wastes at higher pHs. Pure Compounds: Lysozyme (MWabout 13,900) and bovine serum alpumin (MW about,500) occupy nearly the same surface area on activated carbon when adsorbed solution, even though one is 4 times larger (43). These proteins were irreversibly adsorbed. Effective adsorption diffusivities were J about 6 percent of the bulk diffusivity for the lysozyme and 0.7 percent, for the bovine serum albumin. .Maximum adsorption occurred at their isoelectric pOints. The equilibria and kinetics of adsorbing a series of. polyethylene glycols (PEGs, MW 194 to 2,490,000) were studied using coconut-based and lignite-based activated carbons (44). The coconut carbon has most of its pore vol ume in micropores (in this case >70% of the volume is in poressmalier thc;tn 3 nanometers), this does not permit the larger compounds to adsorb as well,with adsorption capacity decreasing for PEGs larger than 1500 MW. With the lignite carbon (>70% of its pore volume is in pores larger than 3 nanometers), adsorption capacity increased with increasing MW up to 1500 MW, where the maximum loading was achieved. Increasing MW above tbis value did not change the adsorption capacity, so the same area or pore volume is accessible for all these large compounds. The effective diffusion coefficient decreased with increasingMW, with the coconut carbon's diffusivities larger than the lignite carbon's. Using these 10

PAGE 16

values in calculations for a "typical" water treatment application found that compounds much smaller than 10,000 MW would pass through the adsorber without having a chance to adsorb, longer contact times are needed. Three different molecular weights of polyvinyl acetate were adsorbed on activated carbon; the smaller compounds took longer to reach "complete" equilibrium; the 170,000 MW molecule took 20 minutes, while the 22,000 MW material took 60 minutes (44). Adsorption equilibria capacities also increased with decreasing molecular size; the 170,000 MW material had an ultimate capacity of 0.06 g/g, the 68,000 MW material's capacity being 0.11 g/g and the 22,000 MW mater falls capacity being 0.19 g/g. Adsorption equilibria were evaluated for different MW fractions of dextran (a: carbohydrate), with adsorption capacity at a maximum for MW of 6000; higher MW fractions (up to 500,000) resulted lower loadings (45). Loadings of 400 mg/g were obtained at equilibria concentrations of 10,000 mg/L. To summar ize, many small compounds (MW < 500) and very large compounds (MW > 20,000) are not very adsorbable, whether they are humic materials or fractionated wastewater components. The most adsorbable fraction is the 1000 to 10,000 MW materials, with up. to 90% of it adsorbable. For many activated carbons, this size material corresponds to the size material that adsorbs to a maximum capacity when adsorbing a homologous series of compounds. The adsorption rate in batch studies decreases with increasing MW, with little or no removal expected in a "typicalfllter adsorber used for water treatment (short contact time); although in column studies all fractions are removed to some degree, even 11

PAGE 17

if they were not removed in batch studies. Solution chemistry is very important for compound adsorption, with calcium and magnesium cations apparently the most important species. 12

PAGE 18

EXPERIMENTAL, Materials: The following chemicals were obtained from either Sigma Chemical, Aldrich Chemical, or Fisher Scientific for use as received: polysaccharides: protein: synthetics: sucrose inulin xylan (342 MW) (5200 MW) (20,000 MW) egg albumin (45,000 MW) phenol (94 MW) polyethylene glycol (400 MW) polyacrylic acid (5,000 MW) polyvinyl pyrrolidone (10,000 MW) The following activated carbons were used as provided by the manufacturers: Filtrasorb 400 Hydrodarco 4000 BAC-SQ Calgon Corporation ICI America Kureha Chemical Co. (Japan) Filtrasorb 400 (F400) is a bituminous'coal.,...based activated carbon with an approximate surface area of 1100 m2/g; Hydrodarco 4000 (HD4000) is a lignite-based carbon with a surface area of 700 m2/g; and BAC-SQ is a petroleum pitch-based carbon with a surface area of 1100 m2/g. The petroleum carbon has the smallest pores, with 67 percent of its pore volume in pores smaller than 2 nanometers (Table I). F400 has 47 percent of its pore volume in pores smaller than 2 nanometers, and HD4000 has only 2 percent of its volume in these small pores, with 60 percent of its volume in pores larger than 10 nanometers. If pore volume is the main factor for the adsorptive capacity of the molecules and surface chemistry is a minor consideration, the largest molecules should be able to penetrate and adsorb better in the HD4000, then F400, 13

PAGE 19

which would be better than BAC-SQi conversely, the smallest molecules should adsorb better in the BAC-SQ, followed by the F400, followed by HD4000. All activated carbon for the equilibria studies was pulverized in a Waring blender, sieved through a u.s. Standard Sieve No. 120 microns), washed with distilled water until the supernatant was clear, dried at 105 C for 24 hours, and stored in sealed bottles until used. TABLE I Pore Volume Distribution for Adsorbents Pore Size nanometers < 2 2 3 3 5 5 -10 > 10 BAC-SQ 67 11 7 6 9 Percent of Total Pore Volume F400 47 22 10 8 13 HD4000 2 12 11 15 60 MilliQ deionized water was used to prepare all for the equilibrium studies using low initial solute concentrations (10 -20 mg/L), the deionized water was filtered through an la-inch deep granular activated carbon bed prior to use, to remove any organic compounds that may be present. No buffer was used since some studies indicate their effects were small (46). Equilibria Studies: Initial equilibria studies were performed using an oscillating platform shaker operating at 150 rpm. Al though the equilibrium time tests were based on this system, the mixing pattern of the activated carbon in solution 14

PAGE 20

did not appear satisfactory, as the activated carbon partially settled in the center of the flask. Wooden racks were constructed and mounted on a metal frame on the oscillating shaker table, the racks held 120 ml Wheaton bottles at a 450 angle -this was done to induce turbulence (Figure 1); also, a drum was modified to hold wooden racks of Wheaton bottles, with the drum then pI aced on a dual roll er dev ice to rotate at low rpms (about 15 rpm was the slowest speed that our system could maintain -Figure 2). Both systems provided better mixing patterns, with no difference noted among the isotherms performed on all three systems. Albumin was used in an initial test to determine the time required for the largest compound (45,000 MW) to reach equilibrium. During the first two hours of shaking, the albumin concentration decreased from 60 mg/L to nearly 30 mg/L, and after 5 days it had dropped to 12.5 mg/L.ln the next 9 days, the c6ncentration decreased to 11 'mg/L, So it appeared 14 days would be sufficient; but to ensure sufficient equilibration, 21 days of mixing were provided. This time is more than adequate based on a comparison to the times used by others (one to seven days) who studied large compound adsorption (23, 25, 28, 30, 36, 38). At the completion of 21 days of mixing, the carbon was settled and then a sufficient quantity of solution was filtered through a Whatman GF/C filter pad; the albumin solution did not filter well and centrifuging rather than filtration to remove the carbon fines frOm solution. Column Studies: Fixed-bed adsorber studies were performed in 25mm internal diameter glass columns of varying 15

PAGE 21

. 0 o o o D II o o o FIGURE I: MODIFIED SHAKER TABLE FOR ISOTHERMS ".\ \ '\ i I ,I :; FIGURE 2: ROLLER APPARATUS FOR ISOTHERMS (\ o o

PAGE 22

lengths to provide the desired contact times. Filtrasorb 400 was used and has an average particle diameter of 1 mm, resulting in a 25:1 column to particle diameter ratio --this is considered by many to be sufficient to eliminate wall effects (47). Masterflex tubing pumps were used to control the flow to the adsorbers, with desired flows of 0.5,1.0,2.0, and 4.0 gpm/ft2. Samples were taken after each column and analyzed for the specific components as indicated below. Analytical Methods: All single component equilibrium samples were analyzed using an Oceanographics International Total Organic Carbon Analyzer Model 525 B. Both the direct inject and the ampule techniques were used; when the final equilib.rium solution concentration was greater than 5 mg/L, the direct inject technique was used, when smaller values resulted, the ampule technique was used. Standard curves were run each time analyses were performed, with the standard prepared using the compound being studied; so results are reported as mg/L as compound and not TOC. When mixtures were used, each mixture contained a polysaccharide (inulin or xylan) which could be determined directly using the phenol-sulfuric acid test (50). Phenol was a second component and ultraviolet spectroscopy (Perkin-Elmer dual beam spectrophotometer Model 5200) was used to determine its concentration. The third component in the mixture was calCulated by measuring the solution's TOC and subtracting the Toe response of the equivalent concentrations of the polysaccharide and phenol and then converting the remaining Toe response into the third compound's concentration. 17

PAGE 23

RESULTS AND DISCUSSION Equilibria Tests: Equilibrium isetherms were perfermed en the three carbehydrates using all three activatedcarbens:F400, HD4000, BAC-SQ. Since F400 is the mest frequently studied adserbent, it was used to. adserb additienal large cempeunds for cemparative purpeses. The Freundl ich adserption equil ibr ium medel was used to. re'present the data, since it has been f6und to. fit many data successfully. Unfertunately it is an empirical medel; but, a theeryhas been prepesed that results in the Freundlich isetherm equatien (49). The theery is based en adserbing substances cn-a surface with heteregeneeus surface energies, therefere'the heat ef adserptien varies with surface ceverage. Th& Freundlich equaticn lin X = K C Where, X is the leading of the inmg/g; C is the equilibrium selutien cencentratien of solute in mg/L; and K and n are empirical K is propertienalte the adserptien energy, with the larger the value,' the greater the adserptien energy; aridn is related to. the intensityef adserptien, with the greater the n, the mere intensely the material is adserbed (49). The Freundl ich isetherm is used to. represent the adserption ef the large cempeunds en the varieus carbens, altheugh its ability to. fit the data was net always geed. The Freundlich censtants are previded in Table II. 18

PAGE 24

TABLE II FREUNDLICH CONSTANTS Compound MW Adsorbent ---------------Sucrose 342 F400 HD4000 BAC-SQ Inulin 5200 F400 HD4000 BAC-SQ Xylan 20,000 F400 HD4000 BAC-SQ ---------------PEG 400 400 F400 HD4000 BAC-SQ PAA 5000 F400 PVP 10,000 F400 Albumin 45,000 F400 Phenol 94 F400 (from reference 46) PEG -polyethylene glycol PAA -polyacrylic acid PVP -polyvinyl pyrrolidone t of data ---------11 9 10 4 7 7 13 5 5 10 6 5 8 6 6 Freundlich Constants K, mg/g n --------------.95 .64 3.03 2.27 7.50 1.66 37.2 4.80 69.6 5.63 84.7 4.35 1.66 2.82 2.07 1.38 1.5E-06 .2.9 40.8 3.89 29.1 4.38 18.8 1.53 6.1E-05 .34 .01 .48 4.5E-04 .35 78.1 4.72 A wide variation in both Freundlich K and n constants exists for the large compounds tested. The adsorption energy term, K, varies by nearly 8 orders of magnitude, with three compounds being essentially non-adsorbable or poorly adsorbed on a particular adsorbent: xylan was non-adsorbable on BAC-SQ; PAA, PVP, and albumin were poorly adsorbed on F400. Inulin and PEG 400 were both adsorbed strongly on all carbons. Adsorption intensity terms varied from .29 to 5.6., with values greater than 19

PAGE 25

1 required for favorable adsorption (49). All three of the additional large compounds adsorbed on F40n were poorly adsorbed, as indicated by the unfavorable n terms. Thexylan was also unfavorably adsorbed on BAC-SQ, it not penetrate its narrow pore structure. In addition, sucrose's n value is less than 1 for F400, but this is misleading since a two-sloped isotherm plot resulted, and these values are for the steep portion of the curve; a horizontal curve results at hlgher concentrations. Sucrose adsorption follows a pattern expected for the adsorption of small molecules based on previous experience (50) and published data using carbons with similar pore structuresf such as a coconut-based (Columbia carbon) or another petroleum carbon (Witco carbons) (5l, 52); the small pore diameter petroleum-based carbon has the greatest adsorptive capacity, followed by the bituminous-based, and, with a considerably lower capacity, the lignite-based carbon (Figure 3). Since the petroleum and bi tuminous coal .carbons have similar surface areas, the main difference is in the pore size distributions (or possibly the surface chemistry which was not evaluated in this study) -the smaller pores apparently exert stronger adsorption forces on the solute, permitting greater adsorption. With the F400 carbon, a two-sloped isotherm results, with the maximum capacity obtained at an equilibrium solution concentration of about 80 mg/L and no additional adsorption occurring with increased concentrations. Multisloped isotherms have been observed by others (53, 54). Sucrose adsorption 20

PAGE 26

"" 60 "-en E ,. C) Z o o ...J x x x x BAC SQ X F400 OHD4000 10 200 !OO SUCROSE CONCENTRATION, FIGURE 3: SUCROSE ISOTHERM

PAGE 27

follows the adsorption pattern expected for smaller compound adsorption. BAC-SQ's adsorptive capacity is nearly double F400's capacity for inulin (MW 5200), a result that is similar to those obtained for sucrose (Fiqure 4). The lignite carbon's capacity is is considerably improved relative to the other carbons, since its capaci ty for the inul in is greater than F400' s and is about thirds of BAC-SQ's (it was about 20% of BAC-SQ's capacity for sucrose)o Inulin's larger size J5200 MW) was expected to limit its access into most of BAC-SQ's pores, and to a large amount of F400's; so the lignite carbon was expected to perform best due to its better performance for the adsorption of humic substances of this size (21). Favorable adsorption behavior is good adsorption capacities for all three activated carbons. The small pores of BAC-SQ are large enough to permit access of the inulin, since. approximately 40% of the total pore volume would be occupied (if' inulin's adsorbed speXflc'gravity'is 1.5). HD4000's pores are better suited to the adsorption of this polysaccharide; the larger moleCUle fills the por'esmore completely, leading to stronger adsorption energies (K constant increased 20 times) and therefore to greater capacities. Xylan (MW does not adsorb on while F400 and 8D4000 capacities are to those with inulin (Figure 5). The large-pore lignite carbon has a much greater capacity than F400. This pattern was expected from the pore size distributions and the results with humic substances (21). With the very large molecules, to obtain better adsorption, a larger pore size distribution is required to obtain 22

PAGE 28

0"\ "-c:rt E .. (!) z 0 60 0 50 ...J 40 J SAC SQ )( F400 o H04000 __ __ ---T 8 8 10 20 50 40 80 80 100 200 INULIN, "",/L FIGURE 4: INULIN ISOTHERMS

PAGE 29

40 30 20 0'\ 10 "-E .. 6 (!) Z o o 4-o -l a 2 2 3 4 6 8 10 XYLAN, FIGURE 5: XYLAN ISOTHERMS )( x x X F400 '.-' .' o HD4000 20 30 40 60 80 100 200 .!: { ,. '-".; m,/L

PAGE 30

the best adsorption capacity. But, its capacity is not large and there may be better means of removing these compounds, such as chemical coagulation or The small, narrow pore size distribution of BAC-SQ was expected to provide excellent adsorptive capacity for the sucrose molecule, with progressively poorer adsorption of the other polysaccharides. Sucrose adsorption is favorable on this carbon, wi th a maximum observed capaci ty of near ly 100 mg/g 1 but the 5200 MW inulin's adsorption is surprisingly good with a maximum capacity of 300 mg/g and a nearly horizontal slope in a Freundlich plot (Figure 6), indicating favorable adsorption. The xylan's adsorption (MW 20,000) is basically non-existent, which is not a, surprise since it should be excluded from this carbon due to the narrow range of pore sizes. Use of this adsorbent finds the molecules of 5200 MW capable of penetrating to pores smaller than 3 nanometers and adsorbing there, with this adsorption stronger than for the smaller molecules1 therefore, this size material should be effectively removed from water and wastewater in studies with all activated carbons. Studies with fractionated'materials found this fraction to adsorbed best (19, 37, 38). Adsorption of the three polysaccharides on F400 exhibits the same pattern, except the xylan adsorbs ,to an appreciable extent (Figure 7). Inulin adsorbs best, with a maximum observed capacity of 105 mg/g and a very shallow slope. Sucrose adsorption follows the Freundlich pattern up to about 80 mg/L equilibrium solution concentration, where its capacity plateaus at 60 70 mg/g. The lI;Yi1an'sads()rption is favorable, 25

PAGE 31

300 INULIN 200 SUCROSE 30 t:rt 20 (!) Z C 10 0 XYLAN ...J e 6 ______ ______ __ 10 20 30 40 60 80 100 200 aoo CONCENTRATION, IL FIGURE 6: ISOTHERMS ON BAC-SQ

PAGE 32

INULIN o o o 0 SUCROSE ,. (!) 20 Z o <{ o ..J XVLAN 20 30 40 60 eo 100 2 0 CONCENTRATION, mg IL FIGURE 7: ISOTHERMS ON F400

PAGE 33

with a low adsorptive capacity (maximum loading about 20 mg/g). F400 has more surface area and pore volume in the pores greater than 2 nanometers than BAC-SQ (53% versus 33%), permitting the larger molecule to penetrate somewhat into the adsorbent. The adsorption pattern for the polysaccharides on HD4000 is slightly different than for other adsorbents (Figure 8). Inulin adsorption is'againthe best adSorbed comp6und, with a maximum observed capacity of l80'mg/g. But, xylan adsorbs better than sucrose on HD4000. This results from a combination of the relatively good adsorption of xylan (maximum loading of 36 mg/g) and the poor -loading of the sucrose (maximum loading of 28 mg/g). The large pore structure of the HD4000 does not provide sufficient pore volume in the micropores for strong adsorption of small molecules, while providing larger pores for stronger xylan adsorption. Unlike the study with polyethylene glycols (44), a maximum capacity was obtained with this 1 igni te carbon, and adsorbing larger similar molecules resulted in, a lower loading; therefore, high loadings are not necessarily expected from the large-pored carbons. The adsorption of additional large compounds was evaluated on F400: PEG -polyethylene glycol 400(MW 400), PAA : polyacrylic acid (MW 5000), PVP -polyvinyl (MW 10,000), and albumin (MW 45,000) (Figure 9). As with the xylan on BAC-SQ, the albumin was poorly adsorbed, with a vertical' isotherm near the starting solution concentration1 its large size prevented any appreciable adsorption. The PEG is favorably' adsorbed, with a Freundlich n of 3.89; adsorptive capacity at 100 mg/L equilibrium concentration was about 130 mg/g or considerably 28;

PAGE 34

200 ... C) Z -o c o ..J ------.-X 3 4. 5 CONCENTRATION I L FIGURE 8 : ISOTHERMS ON HD4000

PAGE 35

o o ...J __ ____ iO 20 30 40 eo 80 100 200 CONCENTRATION, MG/l FIGURE 9: MISCELLANEOUS COMPOUNDS ON F400

PAGE 36

greater than for sucrose's capacity of 66 mg/g at a similar concentration for the similarly sized compounds. So other factors in addition to size 'are important to its adsorption, as is the case for smaller 'compounds frequently studied. Adsorption of the 5000' MW PAA is unfavorable, with an n of .34; surface charge groups on the PAA must be an 'important consideration since this compound is not adsorbed in a manner similar to the comparably sized inulin. The PVP adsorption pattern is also unfavorable (n = .52), al:t:hough the latge compound's adsorption is reasonably good '(64 mg/g) at an equi 1 ibr ium concentration of 60 mg/L. The larger PVP adsorbs better than the PAA, with its capacity somewhat lower than'the smaller inulin molecule, but considerably greater than the larger xylanmolecule. The upper limit of the capacity obtained for the various large compounds tested except for the PAA, indic'atesthat" the optimum size for utilization of theF400 is about 5000, with either decreasing or increas'ingcompound size' tesul ting in lower adsorptive capacities. Note: a limited number of compounds from different families were studied, and very large concentrations not tested to determine the actuailimiting adsorptive A more extensive study of large compound adsorption is needed. Column Tests: The column studies were' intended to monitor the breakthrough pattern of three different molecular weight materials to if size within an *dsorber. If molecular weight site fractionation occurred, it was to be determined if the smaller contpound is able to displace the larger 31

PAGE 37

compounds andresul t.in chromatographi.ceffects" similar to those observed when small compounds .for adsorption '., '. si tes (55). The .inul in, a polysaccharide that can be me.asured by th.e. (48,), phenol, which could be measured by ul traviolet and polyethylene glycol (eitber 1000 or 14,000 molecular weight), .' r; ,. ", .'. '. which would be determined by t,'otal c,arbon -the total orga.nic .' .' ., carbon response of the equivalent amount of phenol and, inul,in would be .subtracted from the sample's TOC, with, the rewaining TOC converted to PEG. Each compound was ,:present at.75 mg!L. Four flows (0.5, ]" 2, & 4 gpm/ft2) were to be uS,edto determine the c '", ; .' ',-.' superf icial veloci ty and/or contc;tct time effect on the breakthrough pattern. While the program appeared, reasc;mableand given insight into large compound adsorption, sev;eral problems developed. First, the j,nvestigat,or. left. the University o,fFlorida in the middle oft.he pr.oject and the .' remaining funds coulqnot be t:ransfered to his new univereity. Second, the graduate student working o.n the, equilibrium portion, was replaced for the column studies, and a student was employed who was not dependent upon theetudy results for his graduation; the principal investigator therefore had little Control oVer the remainder ,of the proj,ect. ThiJ;d.l several problems were encountered wi tb the total a,na,lyzer .andthe backup total organic .carbon ,analyzer, (Beckman 914) and the funds were '., .. ,,' -. not available to. repair .ei the.r ; Fourth when the I diff icul ty with. the Toe apalysis:. the saIllples we,re

PAGE 38

not shipped, as requested, to the principal investigator's new university so the Toe analyses could be performed (too much time may have transpired anyway). Therefore, the results presented here do not include three component interactions within the column, since only two were analyzed. The breakthrough pattern for phenol and inulin at 1 gpm/ft2 flow with PEG 1000 indicates that inulin breaks through before phenol for the first and second columns (empty bed contact times of 7.5, and 15 minutes) (Figure 10); this could be predicted from Freundlich K values, since phenol's K is 78 and inulin's is 37 mg/g (55). But, when the phenol reached the third col umn, 1 i ttl e capacity remained for it, while the inulin continued to be removed (contact time may have a significant effect on adsorption results, but more data are needed). day the inulin broke through, a very steep breakthrough pattern was obtained, with its effluent concentration exceeding phenol's wi thin two days. After 16.' days, the fourth bed was sti 11 removing nearly all the phenol, with a relatively constant leakage of 3 to 4 mg/L of inulin. The PEG 1000 was expected to adsorb well, with its capacity 1 ikely to' be greater than phenol's. How it breaks through the columns is not known, and its effect on the phenol's reversal of order in column three is not known; premature exhaustion may be occur.d ng, but similar results would have been expected in the fourth column. The 1 gpm/ft2 run using the PEG 14,000 results in inulin breaking through prior to the inulin in all three columns (Figure 11); the fourth column had to be removed froIDoperation early in 33

PAGE 39

70 ..J ...... 60 (!) III .. 50 I Z 0 I l-I <{ 40 I OC I l-I Z 30 -15-1 W () I Z 20 I 0 () 10 0 0 2 4 6 B FIGURE BREAKTHROUGH TIME CURVES 10 0 I ..J 60 ...... I C!) I ::i I 50 4 Z I 040 -7.5-1 I l-I <{ I I 30 I I I -15-, Z I I W 020 I I Z I 0 0 10 o 2 4 8 I I .. -,-I NUll N -PHENOL PEG 1000 F 400 CONTACT TIME /53 MIN 10 DAYS FOR 12 14 IS 2 I GPM 1FT '.' ., PEG 14,000 F400 -INULIN 18 20 .PEGIOOO SYSTEM --PHENOL I -30 MIN -,0 CONTACT TIME 10 12 14 I I o I I 16 I P I 18 20 TI ME D.AYS FIGURE II: BREAKTHROUGH 2 CURVES FOR I GPM I FT SYSTEM

PAGE 40

the run due to a rapid headloss buildup. The carbon's capacity for PEG 14,000 is expected to be less than for the phenol and inulin due to its larger size; whether it did remains to be Times to breakthrough for both phenol and inulin are given for the two systems (Table III). PEG size had no effect on inul in breakthroug'h, even though the one PEG was smaller (MW = 1000; inulin = 5200) and.the other larger (MW = 14,000). TABLE III Times to Breakpoint 1 gpm/ft2 System Compound Breakpoint Contact Time, days mg/L Time, min PEG 1000 PEG 14,000 --------------------------------------------Inulin 10 7.5 .7 .5 15 2.8 2.2 30 7.9 8.3 20 7.5 1.0 .7 15 3.4 3.5 30 8.3 8.9 30 7.5 1.2 1.0 15 3.7 3.8 30 9.2 9.5 phenol 10 7.5 2.2 2.1 15 4.9 5.8 30 5.2 12.2 20 7.5 2.6 2.6 15 5.9 6.3 30 7.3 14.7 30 7.5 2.9 2.8 15 6.3 6.7 30 8.3 15.7 This indicates that both PEGs are better or more poorly adsorbed than the inulin, since the inulin apparently is seeing the same number of adsorption sites.within the columns. With phenol, a 35

PAGE 41

difference is noted in the breakthrough pattern of. the phenol. At the greater contact times, the phenol competes better for adsorption sites in the system with the higher MW PEG; this indicates that the larger PEG is not adsorbing as well as the smaller one at deeper locations in the adsorber. When the phenol reaches these points the adsorbent still has plenty of adsorption si tes left for the phenol. Or cooiversely, all the larqer PEG could be adsorbing on the same carbon already loaded with phenol and then the phenol doesn't have to compete for adsorption sites with the smaller PEG for sites: the first hypothesis is more likely due to kinetic considerations. The same two chemical systems were evaluated at gpm/ft2 (Figures 11 & 12). Again, the inulin breaks through the adsorbers before the phenol, in fact it has broken through the 7.5 minute empty bed contact time column before the phenol has broken through the 3.75 minute columns. Breakthrough for both the phenol and the inulin is rapid, with less than a day needed from the time breakthrough begins. The times to different breakpoints were determined for the different PEGs (Table IV). The different PEGs have no apparent effect on either phenol or inulin breakthrough at this higher velocity, and longer contact times were not tested. At 4 gpm/ft2, the headloss was too great for our system and within three days the runs were abandoned with the 7.5 minute contact time systems performing the same for both PEGs. The 0.5 gpm/ft2 runs were not performed, and therefore the longer contact times were not tested. The greatest effects of flow rate were expected to be observed at the slowest flow and greatest contact 36

PAGE 42

70 10 /3.75 I I I I I I 7.5, I o I I I I I I I ,I IS-MIN. CONTACT TIME I I p I I I I I I I / -INULIN --PHENOL PEG 1000 F400 0-4 ____ o 2 4 5 1 9 10 TIME, DAYS FIGURE 12: BREAKTHROUGH CURVES FOR 2 GPM/FT2 SYSTEM, PEG 1000 10 60 (!) ::E ,. 50 Z o 40 0:: I-Z 30 W o Z 020 o 10 I' ./ ;' ;' /3.75 15 t-AI N CONTACT TIME -..,., I / I I' .. ,,1 \ I 1/ \ I G_" I -INUliN --PHENOL PEG 14,000 F400 -. 1-.--e--._ 4P--. o 2 '3 4 5 8 7 8 9 10 TIME, .DAYS FIGURE 13: BREAKTHROUGH CURVES FOR 2 GPM/FT2 SYSTEM, PEG 14000

PAGE 43

time, since it was under similar conditions that premature breakthrough was observed (40). Compound Inulin phenol Breakpoint mg/L 10 20 30 10 20 30 TABJ;..E IV Times to Breakpoint 2 gpm/ft2 system Contact Breakthrough Time, days Time, min PEG 1000, 14,000 -----------_._---------------15 3.4' 3.1 15 3.8 3.3 5 1.1' .6 15 4.1 3.4 3.75 1.1 .9 7.5' 2.6 3.0 15 6.6 6.4 3.75 ,1.3 .9 7.5 3.1 2.9 15 6.8 6.8 3.75 1.5 1.6 7.5 3.3 3.6 15 7.1 7.1 Activated carbon use rates were calculated to the 30 mg/L breakpoint of both phenol and inulin for the several contact times at each flow (Figure 14). With inUlin, little difference in carbon usage is observed between the PEG 1000 (solid line) and the 14,000 (dashed line) at all contact times when the flow is 1 gpm/ft2; but, when the flow is 2 gpm/ft2, a substantial difference was observed at the 7.5 minute contact time. The larger PEG caused the inulin breakthrough much faster, increasing the carbon usage; whether the inulin moved with the PEG or the PEG blocked the adsorption sites is unknown. Phenol 38

PAGE 44

25 (I o o 220 .. ID .J ,. I&J IS (!). <[ en :l CD a:: <[ o 5 PHENOL \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ , \ \ \ \ \ \ \ \ -PEG 1000 --PEG 14,000 INULIN ------.. ----_ _ -----. 02 GPN IFT2 I .PM I FT2 o s 10 IS 20 21 10 CONTACT TIME, MINUTES FIGURE 14: CARBON USAGE: TO A 30 MG/L BREAKPOINT

PAGE 45

breakthrough is unaffected by the PEG at the 2 gpm/ft2, but at the 30 minute contact time with the I gpm/ft2, premature breakthrough is possibly observed with the. smaller PEG. The ." smaller PEG probably proceeded deeper into the adsorber and blocked the pores as hypothesized in the premature breakthrough articl e (19). More contact time need to be performed as well as determination of the PEG breakthrough. Contact time has a significant ,ffect on the carbon usage for the larger inulin molecule, with contact times less 15 minutes causing substantially carbori usages (Figurel4). The carbon usage for phenol breakthrough1s not affected considerably by contact ",except in case for the small,er tl,' PEG at the long contact time. So in case of one molecule, longer contact times are needed f'o,I:' better removal, whilefor'the smaller molecule this may pr;ov;e'>-:'detrimental >l." CONCLUSIONS .. 1. The 5200 MW adsoi:bed much petter .. ; ,'" .. ,', on all three adsorbents than did the 342 MW sucro,se or the 2,0,,000 MW
PAGE 46

4. No chromatographic effect was noted for phenol or inul in in column runs at 1 or 2 gpm/ft2 with either PEG. 5. Inulin breaks through the column before phenol at both flows, as would be expected from the Freundlich K values. 6. Carbon use decreases with increased contact time for inulin in both flow systems. 7. Carbon use decreases with increased contact time for phenol at 2 gpm/ft2 for both PEG systems. 8. Premature exhaustion may occur at longer contact times (30 minutes) at 1 gpm/ft2, since carbon use increased for the PEG 1000 system. 41

PAGE 47

REFERENCES ",. '., o. Nelson" A.J. "Removal. ot,:Large. Weight Organic Compounds by Activated Carbon from Dilute Aqueous Solution", University of Fl,or ida,. .of, Environmental Engineeiing Sciences, Dec. 1982 1. Committee Report,' V.L. snoeyink and Control of Organic Contaminants by Ut;:il itiE:s" Water Works Assn., .&9, 5, 267-271, 1977 2. Hager, D.G. "Activated Carbon'Used for Large Scale Water TreatmentI' Environ .. ScL & Technol.,J.c, 4,,; 287-291., 1967 3. Wood, P.R.; DeMarco, J.; "Chapter 5::R-emovil1g'rotal Organic Carbon and Trihalomethane Precusor Substances" Carbon of Organics Fba.ru;!, Vol. 1, (M.J. McGuire and I.H. Suffet editors) f Ann Arbor SCTence, Ann Arbor, MI., pp. 115 135',1980 4. Brodtmann, N.V., Jr.; DeMarco,J.; GreenOerg,p. IICbapter 8: Critical Study of Large-Scale Granular Activated Carbon Filter Units for the Removal of Organic Substances from Drinking Water" Ibid., pp. 179 -222 5. Yohe, T.C.; Suffet, I.He; Cairo, P.R. "Specific Organic Removal by Granular Activated Carbon Pilot Contactors" Am. WaterWorks Assn., 73, 8, 402-410, 1981 6. Environ. Protection Agency, "Control of Organic Contaminants in Drinking Water" Federal Register, 43, 28, 5756-5758, 1978 7. Pendygraft, G.W.; et al "The EPA-Proposed Granular Activated Carbon Treatment Requirement: Panacea or Pandora's Box?" .Journ, Am. Water Works Assn., 11, 2, 52-60, 1979 8. O'Brein, R.P.; Jordan, D.M.; Musser, W.R. "Trace Organic Removal from Contaminated Groundwaters with Granular Activated Carbon" Paper, Am. Chern. Soc. Mtg., 3/29 4/3/81, Atlanta, GA 9. Anon, "The Calgon Adsorption Service" Calgon Corp. Brochure 10. Slechta, A.F.i Culp, G.L. "Water Reclamation Studies at the South Lake Tahoe Public Utility District" Journ. Water Pollution Control Fed., 39, 5,787-814, 1967 11. McCarty, P.L.; et al "Water Factory 21: Reclaimed Water, Volatile Organics, Virus, and Treatment Performance" EPA=-_ 600/2-78-076, Env iron. Protect. Agency, Washington, D.Co June 1978 12. Weber, W.J., Jr.; Hopkins, C.B.; Bloom, "Physicochemical Treatment of wastewater" Journ. Water Pollution Control Fe,d., 1, 83-99, 1970 42

PAGE 48

13. Hager, D.G. "A Survey of Industrial wastewater Treatment by Granular Activated Carbon -Part II", Paper at 68th Annual Am. Instit. of Engrs. Mtg., Nov. '1975, Los Angeles, CA 14. Environ. Protection Agency, "Preliminary Assessment of Suspected Carcinogens in water" Report to U.S. Congress Office of Toxic Substances, Washington, D.C., Dec. 1975 15. National Academy of Sciences, "Drinking water and Health", Safe Drinking Water Committee, National Research Council, Washington, D.C., 1977 16. Black, A.P.; Christman, R.F. "Chemical Characteristics of Ful vic Acids" Journ. Am. Water Works Assn., 55, 7, 897-912, 1963 17. Gjessing, E.; Lee, G.F. "Fractionation of Organic Matter in Natural Waters on Sephadex Columns" Environ. Sci. & Technol., 1:, 8, 631-638, 1967 18. Schoor, J.L.; et al "Trihalomethane Yields as a Function of PreCursor Molecular Weight" Environ. Sci. & Technol., 13, 9, 1134-1138, 1979 19. Arbuckle, W.B. "Chapter 10: Premature Exhaustion of Activated Carbon Columns" ,Activated Carbon Adsorption of Organics from the Aqueous Phase, Vol. 2, (M.J. McGui re and I.H. Suffet editors), Ann Arbor Science, Ann Arbor, MI.,pp.237-252,1980 20. Davis, J.A.; Gloor, R. "Adsorption of Dissolved Organics in Lake water by Aluminum Oxide. Effect of Molecular Weight" Sci. & Technol., 15, 10, 1223-1229, 1981 21. Lee, M.C.; Snoeyink, V.L.; Crittenden; J.C. "Activated Carbon Adsorption of Humic Substances" J.Arn. Water Works Assn., 73, 8, 440-446, 1981 22. Manos, G.P. ; Tsai ,C-E "Mechanisms of Humic Material Adsorption on Kaolin Clay and Activated Carbon" Water, Air, Soi!. Pollutio.!1, 14, 419...;427, 1980 23. McCreary, J.J.; SrlOeyink, V.L. "Characterization and Activated Carbon Adsorption of Several Humic Substances" Water Researcl:t, 14, 2, 151-160, 198Q 24. Boening, P.H.; Beckmann, D.D.; Snoey ink, V.L. "Aci tv a ted Carbon versus Resin Adsorption of Humic Substances" Journ Arn. Water Works Assn., 72, 1, 54-59, 1980 25. Weber, W.J., Jr.; et al "Chapter 16: Potential Mechanisms for Removal of Humic Acids from Water by Activated Carbon" Activated Carbon Adsorption of Organics from the Aqueous Phase, Vol. 1 (I.H. Suffect & M.J. McGuire editors), Ann Arbor Science, Ann Arbor, MI., 317-336, 1980 43

PAGE 49

26. Randtke,' S.J.; "Effects ,of Sa,l ts ,on A,cti vated Carbon Adsorption of Fulvic,Acid:S" Journ.Am. Assn., 74, 2, 84-93, 1982 ," ,'. I, 27. Randtke, Jepsen, Pretreatmen:t,' 'tor ,J. ",Am Water Works,Assn,]3, 8, 411-419, 1961 28. Youssefi, M.; Faust, S.D. 5,:. AdsQrptio"n, and Formation of Light Halogenated Hyqrocarpons and',Humi,c Acid in Water by Granul ar Acti vated Carbon" Activated Carbon Adsorption of Organics from the Aqueous, Phase,_ Vol. ,1, (:Ii.iH. Suffet & M.J; McGuireeditors),'Anri A,i:porS,cience, Ann Arbor, MI., 133-143,1980 ,,' ,'" 29. Murin, C.J.; Snoeyink, V.L. nCompeti ti ve of 2,4-dichlorophenoland 2,4,6-trichlorophenol in' the Nanom.olar to Micromolar Concentration: Ran'ge" Environ. 13, 3, 305-311, 1979 30. Huang, J-C; Garrett, J.T., "Effects of Colloidal Materials and Polyelectrolytes Carbon in Aqueous Solutionn Proc. Purdue Industrial waste Conf., 30th, Ann Arbor Science, Ann Arbor, 1111-1121, 1976, 31. Ward, T.M.; Getzen, F.w."Influence of pHon the Adso!ption "of Aromatic Acids on Ac'tivated Carbon":, ,Envi,ro'n: Sci.' & Technol. ,-.!, i, 1970 ; 32. Lee, M.C.; et al nDesign' of t.p Remove ,Humic Substances" Environ. Engr,. Soc. of Civil Engrs., 109, EE3,' 631-645, 1983 33. Zuckerman, M.M.; Molof, nHighQualityReUf3eWater by Chemical-Physical Wastewater TreatJllent '!' Journ. Pollution Control Fed., 42, 3,437-456, 1970 34. Guiruis, W.;' et al "Improved Carbon by Pre-ozonation" Poll ution ,50" 2, 308-320, 1978 35. Gerba, C.P.;et "Adsorption of onto. Activated Carbon iri Wastewater" Environ. Sci. & 'l'echnol.,9, 8,' 727-731, 1975 --36. Summers,R.S.; of DOC Rernov,al in Activated Carbon 'Beds" paper tted,to Div., Am Soc. Civil Engrs., 1983 37. Kim,B.R.; Snoeyink, V.L.; Saun'derEi,' of Activated Sludge Cell Time, on Environ. Engr. Div., Am Soc. Civil Engrs., 102"EEl" 55-70, 1976 ", '. 44

PAGE 50

38. DeWalle, F.B.; Cbian, E.S.K. "Removal of Organic Matter by Activated Carbon Columns" J. Environ. Engr. Div., Civil Engrs., 100, EE5, 1089-1104, 1974 39. Chow, O.K.; David, M.M. "Compounds Resistant to Carbon Adsorption in Municipal Wastewater Treatment" J. Am. Water 69, 10, 555-561, 1977 40. Sherm, M.; Lawson, C.T. "Activated Carbon Adsorption of Organic Chemical Manufacturing Wastewaters After Extensive Biological Treatment" ___ Am. Inst. Chern. Engrs./Environ. Protection Agency, pp. 196 203, 1976 41. Volesky, Roy, C. "The Effect of Polysaccharidic Gums on Activated Carbon Treatment of Textile Wastewaters" Water Research, 13, 8, 791-800, 1979 42. Wang, L.K.; et al "Adsorption of Dissolved Organics from Industrial Effluents onto Activated Carbon" Journ. of Ghem. Biotechnol., _25, 491-502, 1975 43. Wu, R-Y.A.; Okos, M.R. "Adsorption of Proteins onto Acti va ted Carbon and :Phenol ic Resins n paper 1982 November 1982, Los Angles, CA. 44. Suzuki, M.; Kawai, T.; Kawaqoe, K. "Adsorption of Poly (oxyethylene)of Various Molecular Weights from Aqueous Solutions on Activated Carbon" __ 3, 203-208, 1976 44. Claesson, I.; Claesson, S. "The Adsorption of Some High Molecular Substances on Activated Carbon" Arkiv for Kemi, r __ ., 19, 5, 1-12 1 944 45. Ingelman, Be; Halling, M.S. "Some Physico-chemical Experiments on Fractions of Dextran" Arkiv for Kemi,l, 10, 61-80, 1949. 46. Peel, R.G.; Benedek, A. "Attainment of Equilibrium in Activated Carbon Isotherm Studies" __ Technol., 14, 1, 66-71, 1980 47. Gilliland, E.R.; Baddour, R.F. "The Rate of Ion Exchange" Ind .. __ __ 2, 330-337, 1953 48. Benef iel d, L.D.; Randall, C.W. "The Phenol-SuI fur ic Acid Test: Effective Alternative for Carbohydrate Analysis" Water __ 2, 55, 1976 --------49. Adamson, "Chapter IX: The Solid-'Liquid Interface Adsorption from Solution" C_hemisty of Surfaces, __ Wiley-Inter science, New York, 386-394, 1976 -50. Arbuckle, W.B. Unpublished reports Union Carbide Corp. 45

PAGE 51

51. Zagorski, J.S.; Faust, S.D. nRemoval of Polluted Watersn water June 1974' .. 52. Mullins, R.L., Jr.; etal nThe of Several Brands of Granular ActivatedCarpon for the Removal of Tr ihal omethanes f rom Dr inking water" v ated Carbon Adsorption _.9f_ OrgaIl!CS from .tlHLAqy.e.Q.u.s..l:h.a.B.e, VQ1..,.l, (I.H. Suffet & M.J. McGuire editors), Ann Arbor ScL, Ann Arbor, MI, pp.273-307 1980' 53. Snoeyink, V.L.; Weber, W.J., Jr.; Mark, Jr. "Sorption of Phenol and Nitrophenol by Activated Carbonn .E:nv,irQI1o.$gi. 10, 918-926, 1969 54. AI-Bahrani, K.S.; Martin, R.J. "Adsorption Studies Using Gas-Liquid Chromatography -I. Effect of Molecular Structure" Water Researgh, _10, 731-:-736, 1976 55. Arbuckle, W.B.; Romagnol i, R.J. nprediction of the Preferentially Adsorbed Compound in Bisolute Column Studies", Chern. Engr. Prog. Symp. Sere Water"719.79, 76, 197,61-71, 1980 46