1-s2.0-0160412092900272-main.pdf

Environment International, Vol. 18, pp. 597-607, 1992 Printed in the U.S.A. All rights reserved.

0160-4120/92 $5.00 +.00 Copyright©1992 Pergamon Press Ltd.

QUANTITATIVE EVALUATION OF XAD-8 AND XAD-4 RESINS USED IN TANDEM FOR REMOVING ORGANIC SOLUTES FROM WATER

Ronald L. Malcolm U.S. Geological Survey, Box 25046, Mail Stop 408, Denver, CO 80002, USA

Patrick MacCarthy Department of Chemistry and Geochemistry, Colorado School of Mines, Golden, CO 60401, USA

El 9207-165a M (Received 12 July 1992; accepted I September 1992)

The combined XAD-g and XAD-4 resin procedure for the isolation of dissolved organic solutes from water was found to isolate 85% or more of the organic solutes from Lake Skjervatjern in Norway. Approximately 65% of the dissolved organic carbon (DEC) was first removed on XAD-8 resin, and then an additional 20% of the D e c was removed on XAD-4 resin. Approximately 15% of the D e c solutes (primarily hydrophilic neutrals) were not sorbed or concentrated by the procedure. Of the 65% of the solutes removed on XAD-8 resin, 40% were fulvic acids, 16% were humio acids, and 9% were hydrophobic neutrals. Approximately 20% of the hydrophilic solutes that pass through the XAD-8 resin were sorbed solutes on the second resin, XAD-4 (i.e., they were hydrophobic relative to the XAD-4 resin). The fraction sorbed on XAD-4 resin was called XAD-4 acids because it represented approximately 85-90% of the hydrophilic XAD-8 acid fraction according to the original XAD-8 fractionation procedure. The recovery of hydrophobic acids (fulvic acids and humic acids) and the hydrophobio neutral fraction from XAD-8 resin was essentially quantitative at 96%, 98%, and 86%, respectively. The recovery of XAD-4 acids from the XAD-4 resin was only about 50%. The exact reason for this moderately low recovery is unknown, but could result from /¢-g bonding between these organic solutes and the aromatic matrix of XAD-4. The hydrophobic/hydrophilic solute separation on XAD-8 resin for water from background Side A and Side B of the lake was almost identical at 65 and 67%, respectively. This result suggested that both sides of the lake are similar in organic chemical composition even though the D e c variation from side to side is 20%.

INTRODUCTION and the subsequent elution of these solutes from the resins. It is well documented that anion exchange resins commonly remove large quantities of organic solutes from water (Packham 1964), but the recovery of the organic solutes from the resin is usually extremely low.

Several different resin techniques have been proposed and used for the isolation of organic solutes from water. Although organic isolates are obtained from the different procedures, only a few studies have been conducted to quantitatively investigate the sorption of organic solutes from water by various resins 597

598

In recent years, various XAD resins have been used to isolate organic solutes (Mantoura and Riley 1975; Malcolm et al 1977; Thurman and Malcolm 1981; Leenheer 1981) from water. Because the properties of most of the XAD resins are well understood and extensive studies have been conducted to evaluate resin performance in the isolation and recovery of organic solutes from water (Aiken et al. 1979; Thurman et al. 1978a; Thurman et al 1978b), the XAD-8 resin procedure has been extensively used in water chemistry research. It has been established that the so-called hydrophobic acid fraction (i.e. fulvic and humic acids) and the hydrophobic neutral fraction can be quantitatively concentrated from water by this resin and then quantitatively recovered from the resin. These hydrophobic components commonly represent 50 to 65% of the dissolved organic carbon (DOC) in water (Malcolm 1985). Another large fraction of the organic solutes in water, the so-called hydrophilic acids which will not sorb on XAD-8 resin, commonly represents 20-30% of the DOC (Malcolm 1985). This hydrophilic acid fraction has been isolated from water by use of the weak-base anion-exchange resin, Duolite A-7. Although this method has been used with some success, the recovery of hydrophilic acids from the anion-exchange resin is not well documented. Also, because inorganic anions are concentrated along with the organic anions, the procedure becomes less effective with increasing salinity or conductivity of the waters. An alternative to using Duolite A-7 (Leenheer and Noyes 1984) for the isolation of hydrophilic acids from water is the use of XAD-4 resins following XAD-8 in the isolation scheme. This method also has been used without adequate quantification. But the combined use of XAD-8 and XAD-4 resins appeared to be promising and the tandem technique was selected for use in the isolation of organic solutes in the HUMEX acid rain project in Norway. This important environmental study requires accurate and precise methods to detect possible slow, long-term changes in the environment; therefore, the quantitative evaluation of the XAD-8/XAD-4 method was necessary as an integral part of the study. Accordingly, the purposes of this paper are: 1) to discuss the theory and practice of the combined XAD-8 and XAD-4 isolation procedure, 2) to present the removal and recovery data for the isolates of the procedure, and 3) to discuss factors which may affect the composition and character of the organic isolates obtained by this procedure.

R.L. Malcolm and P. MacCarthy

METHODS AND MATERIALS The site

The DOC source used in this study was Lake Skjervatjern which is in the acid rain (HUMEX) study site near Ferde, Norway. This lake has been selected for a comprehensive study of the effects of acid rain. The lake had been divided into two parts by a thick plastic curtain extending from floats on the surface of the lake to the bottom sediments. One part of this lake and associated watershed were acidified by adding sulfuric acid, while the other part was not acidified. The organic solutes in the lake on both sides of the dividing curtain will be studied over a period of years in order to determine what effects the acidification will have on the nature of the organic solutes. This paper presents the results of some of the background DOC studies on both sides of the lake prior to the start of acidification in October 1990. Side B will remain as unacidified, background, or Control Side B for the entire term of the acid-rain experiment. Side A will be referred in this paper as Background Side A because at the time of organic sampling for this paper (September 1990), it was not yet acidified and the DOC data represented background data for Side A. For more details and information concerning the study site, refer to the Humor/Humex Newsletter published by the Norwegian Institute for Water Research (NIVA), Oslo (Gjessing 1992).

Water sampling Water samples from Lake Skjervatjern for this study were taken on September 19, 1990, prior to the first acidification treatment of Side A during October 1990. Flowing water was sampled from lake spillways on both Side A and Side B of the lake. Approximately 1300 L of water from each side of the lake were collected in 37-L, stainless-steel cans, transported the same day by truck to the NIVA laboratory in Oslo, pressure-filtered through 293 Ixm Gelman membrane filters of 0.45 lxm pore size without any wetting agents. The filtrate was collected in 45-L glass jugs, and then acidified to pH 1.95 with concentrated HCI.

Resin isolation procedure The isolation procedure used two types of XAD resin in series: XAD-8 resin first, followed by XAD4 resin. The acidified, filtered water from each side of the lake was first passed through a 4-L column of XAD-8 resin. The effluent from the XAD-8 column was then passed through a 4-L column of XAD-4

Removal of organic solutes from water by resins

resin. Fulvic acids, humic acids, and the hydrophobic neutral fraction were quantitatively sorbed onto XAD-8 resin according to the procedure of Malcolm (1991). For a k0.5 = 50, 135 L of sample were passed through the XAD-8 column before elution. This 135-L portion of sample was called a "run". After each run, the fulvic acids and humic acids (the so-called hydrophobic acid fraction) were back-eluted from the XAD-8 with dilute 0.1 M NAOH using all the precautions mentioned in Malcolm (1991). The eluate was immediately acidified with dilute HC1. The XAD-8 column was regenerated with acid and the next run initiated. After elution of fulvic and humic acids from the final run, the XAD-8 resin was acidified to pH 2 with dilute HC1, stored in glass containers, and transported to the U.S. Geological Research L a b o r a t o r y in Arvada, CO, U.S.A. The resin was then washed with distilled water until chloride free, placed in a 2-L Soxhlet extraction thimble, and the accumulated hydrophobic neutral solutes were extracted using a large Soxhlet apparatus for two days in an extractant mixture of acetonitrite (750 mL) and water (250 mL). The extract was evaporated to near dryness, water was added, and the sample was evaporated to near dryness again in a vacuum concentrator at room temperature. The concentrate was then freeze-dried. The effluent from the XAD-8 column (i.e., the solution that passed through the column at low pH) was pumped into a 45-L glass reservoir. This effluent was then pumped onto a 4-L column of XAD-4 resin. The XAD-4 resin was eluted with dilute base after each 135-L run in the same manner as for the XAD-8 resin. The XAD-4 resin was acidified to pH 2.0 after each such elution and the next run initiated. The solutes eluted from the XAD-4 resin by dilute base are referred to as the XAD-4 acids. The eluates from the XAD-8 and XAD-4 resins were stored in glass containers and then transferred to plastic containers for transportation to the research laboratory in Arvada, CO. The eluate from the XAD8 resin was separated into fulvic and humic acid fractions at pH 1.0. All samples of humic acid, fulvic acid, and XAD-4 acids were reconcentrated separately onto the resin on which they were isolated, then de-salted, eluted with dilute base, hydrogen-saturated by cation exchange, vacuum concentrated at room temperature and, lastly, freeze-dried according to the methods of Malcolm (1991). The XAD-8 and XAD-4 column effluents were sampled for DOC analysis after 22.5 L, 45.0 L, 67.5 L, 90.0 L, 112.5 L, and 135 L of each of the two water samples had passed through the two different resin

599

columns during Runs 3 and 5. For both resins, the column void volume (same as column pore volume) is 67% of the resin bed volume (the void volume for a 4-L column of resin is 2.68 L). Therefore, the column effluent volumes of 22.5 L, 45 L, 67 L, 90 L, 112.5 L, and 135 L are equivalent to 8.4, 16.8, 25.0, 33.6, 42.0, and 50.0 pore volumes of sample passing through the 4-L column.

DOC analysis Several samples of the filtered water from each side of the lake were collected in 10-mL glass vials which were previously baked at 450 ° C to remove all organic contaminants. The vials were sealed with Teflon screw-on caps and immersed in an ice-bath until analysis in Denver, CO three weeks later. Samples of column effluents were also collected and stored in the same manner. DOC analysis was conducted on duplicate 3-mL samples using an automated Oceanographics International Carbon Analyzer. Sample standards were run after each ten samples. The carbon analyzer was standardized for a DOC range from 0.1 mg C/L to 10 mg C/L. Sample accuracy was +0.03 mg C/L using potassium acid phthalate and fulvic acid standards. Sample precision was (standard deviation) +0.02 mg C/L.

Elemental analysis Elemental analyses (C, H, O, N, S, P), and moisture and ash determinations were conducted by Huffman Laboratories, Inc., Golden, CO. THEORETICAL AND PRACTICAL CONSIDERATIONS RELATING TO THE COMBINED RESIN PROCEDURE

The combined resin procedure (XAD-8 followed by XAD-4) is hypothesized to recover a large part of the natural DOC solutes by a simple and uncomplicated method without contaminating the solutes with organic solvents or reagents. The large hydrophobic macromolecules of fulvic acids, humic acids, and the hydrophobic neutral fraction are quantitatively sorbed and then quantitatively recovered from the XAD-8 resin. XAD-8 is the only resin that has been found to quantitatively sorb macromolecules common in natural waters, and from which these solutes can also be quantitatively eluted. The large pore size (25 nm or greater) in this resin facilitates rapid diffusion of macromolecules and solvents into and out of the macroporous network of the resin matrix. After these large molecules are removed by XAD-8 resin sorption, the more hydrophobic XAD-4 resin with a

R.L. Malcolm and E MacCarthy

600

large surface area can be used to sorb a large proportion o f the l o w e r - m o l e c u l a r - w e i g h t organic acids which are not sorbed on XAD-8 resin (i.e., they are hydrophilic relative to XAD-8 resin). The XAD-4 resin matrix o f stryrene divinylbenzene is uncharged like XAD-8 resin, but the XAD-4 resin has essentially no polarity; thus, it is a more hydrophobic resin than the somewhat polar XAD-8 acrylic ester-based resin. The additional acid fraction, XAD-4 acids, which are sorbed by XAD-4 resin, are eluted from this resin with dilute base in the same manner as is the hydrophobic acid fraction (fulvic acids and humic acids) from the XAD-8 resin. The isolation and recovery of this additional fraction (the XAD-4 acids) could increase the r e c o v e r y of dissolved organic solutes from 45-50% on XAD-8 resin alone to 70-85% with both resins. The possible isolation o f 70-85% o f the dissolved organic solutes would be a significant accomplishment for water chemistry studies. The use o f XAD-4 resin in the isolation procedure may surprise some investigators because its use as the first or only resin to isolate solutes from natural

waters has been strongly discouraged (Aiken 1979; Malcolm 1985). It must be emphasized that XAD-4 resin is to be used o n l y in the isolation s c h e m e after the w a t e r has b e e n p r e v i o u s l y p r o c e s s e d t h r o u g h XAD-8 r e s i n to r e m o v e the h y d r o p h o b i c m a c r o molecular solutes. The XAD-4 resin has a large capacity for small, uncharged solutes, but is ineffective when used alone, or as the first resin in a sequence of different resins. The small pores (5 nm or less) o f the highly crossl i n k e d X A D - 4 r e s i n p a r t i a l l y e x c l u d e the l a r g e macromolecules o f fulvic acids, humic acids, and hydrophobic neutral constituents. The resin pores XAD-4 are also blocked with these solutes and this prevents the large internal macroporous surface area o f the resin from being utilized for solute sorption. Because of these factors, the XAD-4 resin is prevented from quantitatively sorbing solutes from natural waters and this results in a low sorption capacity for the resin and a rapid breakthrough of hydrophobic solutes from the resin. It was also found that only about 75% of the small amounts o f hydrophobic solutes sorbed by the XAD-4 resin could be r e c o v e r e d by elution

Table 1. DOC values (mg/L of the XAD-8 column effluent during Runs 3 and 5 of Skjervatjern Lake water for both sides of the lake during September 1990 (Influcnt DOC = 6.70 mg C/L on Side A; Influent DOC = 8.40 mg C/L on Side B). Side A Effluent Volume In Liters

22.5

Effluent VolumeIn ColumnPore Volumes

8.4

Run 3

1.88

Run 5

Run 3

Run 5

2.14

2.27

2.04

1.98 45.0

17.8

2.08

2.24 2.44

2.20 67.5

25.0

2.31

33.6

2.33

2.68

42.0

2.45

2.66

50.0

2.54 2.54

2.55

3.00

2.88

3.06

2.87 2.63

2.52 135.0

2.77

2.64

2.36 112.5

2.51 2.45

2.27 90.0

$i~1~B

2.98

3.13

2.97 2.80

3.05 3.03

3.26

Removal of organic solutes from water by resins

601

from the resin. The low recovery of sorbed solutes was attributed to pore clogging and rc-n interactions between the aromatic matrix of the XAD-4 resin and the aromatic portion of natural organic solutes. RESULTS AND DISCUSSION The DOC values of the filtered water from Background Side A and Background Side B of Lake Skjervatjern were 8.40 mg C/L and 6.70 mg C/L, respectively. These DOC values represent an average of four samples taken from each side of the lake. A total of 1282 L of water from Side B of the lake were processed in 9.5 runs on the resin columns. A total of 1035 L of water from Side A of the lake were processed in 7.7 runs. The DOC values for XAD-8 and XAD-4 column effluents, respectively, during Runs 3 and 5 for Sides A and B o f the lake are given in Tables 1 and 2. The combined XAD-8 and XAD-4 resin system sorbed 85% (5.68/6.70) and 87% (7.23/8.40) of the DOC

from Side A and Side B, respectively. The first resin (XAD-8) removed 67% (5.61/8.4) of the DOC from Side B resulting in a h y d r o p h o b i c / h y d r o p h i l i c separation of 67%/33%. XAD-8 resin removed a similar large proportion of the DOC (64% = 4.28/6.70) from Side Awith a h y d r o p h o b i c / h y d r o p h i l i c se p a ra tio n of 64%/36%. These high percentages of hydrophobic solute removal on XAD-8 are common for waters with high concentrations of fulvic and humic acids. The XAD-4 resin sorbed a high p e r c e n t a g e of the lower-molecular-weight hydrophilic acids which passed through XAD-8 resin. These acids are referred to as XAD-4 acids and account for 21% (5.68/6.70 4.28/6.70) and 20% (7.27/8.40 - 5.61/8.40) of the DOC from Sides A and Side B, respectively. Only 15% (1.02/6.70) and 13% (1.13/8.40) are hydrophilic unsorbed DOC components of Side A and Side B, respectively. Because the hydrophilic neutral fraction is commonly 8-10% of the DOC, and hydrophilic and hydrophobic bases combined represent 2-4% of

Table 2. DOC values in mg C/L of the XAD-4 column effluent during Runs 3 and 5 of XAD-8 processed Skjervatjernlake water for both sides of the lake during September1990 (InfluentDOC 2.42 ms C/L on Side A; Influent DOC 2.79 mg C/L on BackgroundSide B).

Side A Effluent Volume In Liters

22.5

Effluent VolumeIn ColumnPore Volumes

8.4

Run 3

0.75

Run 5

0.78

0.73 45.0

17.8

0.87

25.0

0.94

1.02

33.6

1.11

1.04

42.0

1.09

1.08

50.0

1.14 1.14

1.68

1.05

1.09

1.18

1.15

1.22

1.09 1.22

1.17 135.0

0.88

1.07

0.94 112.5

0.88

Run 5

1.70

0.99 90.0

Run 3

0.87

0.84 67.5

Side B

1.21

1.26

1.16 1.19

1.17 1.21

1.40

R.L. Malcolm and P. MacCarthy

602

Table 3. Isolate weight, carbon content, and recovery data of fulvic acids, humic acids, hydrophobic neutrals, and XAD-4 acids for both sides of Lake Skjervatjem in September 1990.

Isolate

Isolate % C of W ~ . in ~. Isolate

Wgt. of C in Isolate

% DOC of Isolate

% Recovery of Isolate

Background Side A Fulvic Acids

4.682

54.69

2.561

39

95

HnmicAcids

1.923

56.57

1.088

16

98

HPO S ~ , t ~

1.069

50.50

0.542

9

48

XAD-4Acids

1.385

49.90

0.691

21

87

4.182

40

96

Background Side B Fulvie Acids

7.612

54.23

HumicAcids

3.048

55.79

1.700

16

99

HPO t~utr~

1.990

50.70

0.972

11

85

XAD-4

1.935

50.25

1.009

20

45

the DOC, the remaining hydrophilic acid fraction from XAD-4 must be very low. Therefore, the XAD-4 resin is believed to have sorbed 85-90% of the lowermolecular-weight XAD-8 hydrophilic acids. The hydrophobic fraction of XAD-8 contains the fulvic acids, humic acids, and the hydrophobic neutral fraction. After the fulvic and humic acids were eluted with dilute base in the final run, the hydrophobic neutral fraction was extracted with

aqueous acetonitrite (750 mL/L), twice concentrated to near dryness, and then freeze-dried. As shown in Table 3, the combined isolates of fulvic acids, humic acids, and the hydrophobic neutral fraction from Side B of Lake Skjervatjern contained 6.854 g of carbon and represented 67% of the DOC. The respective hydrophobic isolates of fulvic acids, humic acids, and hydrophobic neutrals from Side A represented an

40 "0

_

0 E

3

60 o o •

E

L

o

o

0 r-~

•

2

G

80

0

1

tO)

O_

100

8.4

17.8

25

33.6

42

50

Column pore volumes of sample passed through column Fig. 1. DOC breakthrough curves for Runs 3 ( I ) and 5 (O) Side A water on XAD-8 resin.

Removal of organic solutes from water by resins

603

acids from the XAD-4 resin is unknown, but will be investigated in future experiments. The breakthrough curves for DOC for Runs 3 and 5 of Side A and B for both resins are shown in Figs. 1-4. There are many aspects of these curves which must be discussed. All the breakthrough curves for both XAD-8 and XAD-4 resins correspond to the flat portion of the idealized breakthrough curve, even though in reality they show a gradual increase in DOC breakthrough with increasing volume of water passing through each column. The DOC breakthrough curve on XAD-8 resin is slightly higher for Run 5 than for Run 3 as seen by comparing breakthrough curves in Fig. 2 for Side B

almost identical percentage of the DOC as 39%, 16% and 9%, respectively. Based upon the DOC data, the amount of water processed, weight of hydrophobic neutrals, the actual weight of each isolate, and the C content of each isolate, the percentage recovery for each isolate can be calculated. The recovery data are also presented in Table 3. The recovery of fulvic acids and humic acids from XAD-8 resin is excellent at near complete recovery. The recovery of hydrophobic neutrals from XAD-8 is good at 86%. The recovery of XAD-4 acids from XAD-4 resin is only about 50%, which is not good. The reason for the poor recovery of XAD-4

"0 0

-~ 3

0

•

o

E

2

G 0

0

0

60 ~

0 •

I

0

0

80 u

r~

i i00 8.4 17'.8 25 33~6 42 50 Column pore volumes of sample passed through column Fig. 2. DOC breakthrough curves for Runs 3

(I) and 5 ( 0 ) Side B water on XAD-8 resin.

40 "X3

0

-

3

60

E QJ I0 r~

E .

2

(-

80

0

0

0 •

0 1

0

0

(D (_} £_ t3..

I

I

I00

8.'4

17.8

25

33'.6

42

50

Column pore volumes of sample passed through column Fig, 3. DOC breakthrough curves for Runs 3 (s) and 5 (O) of Side A water on XAD-4 resin.

604

R.L. Malcolm and P. MacCarthy

1

I ° E $-

O O

E

G

I

e-

O

O

O m

O

O •

O •

U

II

8.4

17.8

25

33.6

42

100 50

Column pore volumes of sample passed through column Fig. 4. DOC breakthrough curves for Runs 3 (m) and 5 (O) of Side B water on XAD-4 resin.

water. This effect results from a gradual build-up of hydrophobic neutral solutes on the XAD-8 resin with each additional run on the same column. After each run, the hydrophobic acids (fulvic and humic acids) are eluted from the column with dilute base. The hydrophobic neutral fraction is not removed from run-to-run on the column, but gradually accumulates on the resin resulting in slightly less column capacity or k' with each successive run. Because the hydrophobic neutral fraction commonly represents between 5-10% of the DOC, approximately 5 to 7 runs may be made before the build-up of hydrophobic neutrals has a pronounced effect upon k' If the percentage of hydrophobic neutral solutes is not known, and if there is a possibility that this percentage may be higher than 5-10% of the DOC for a new or given water sample, DOC breakthrough curves should be determined on Run 3 and subsequent runs to establish that a k' below 50 does not occur. When and if a k' of 50 or less is achieved, another batch of clean resin must be used to obtain additional isolates of that water sample in order to maintain quality control over the isoltes. The same phenomenon of the gradual build-up of hydrophobic neutrals affecting DOC breakthrough and k' is exhibited between Run 3 and Run 5 (Fig. 1) on the Side A water sample. The slightly higher last point on the DOC breakthrough curve for Run 5 on the XAD-8 resin for the Side A sample (Fig. 1) may indicate the approaching breakthrough of fulvic acids on that column. Accurate and precise DOC data are essential to clearly define the chromatographic processes, and to

accurately determine the DOC break-through curve, the hydrophobic/hydrophilic separation, and the initial DOC. Precise DOC data are important in establishing DOC break-through curves for samples of initially low DOC and are especially critical in establishing DOC breakthrough curves for the effluent of XAD-4 resins because the DOC values are commonly below 1 mg C/L. The precise DOC breakthrough curves and their interpretation for the XAD-4 resin runs (Figs. 3 and 4) cannot be obtained without exacting methods of DOC analysis. The breakthrough curves for Runs 3 and 5 on XAD-4 for the same sample (Fig. 3 or Fig. 4) are essentially the same and do not show any solute build-up on the resin with sequential runs.

COMPOSITION OF XAD-4 ACID FRACTIONS

The breakthrough curves for hydrophobic solutes (fulvic acid, humic acid, and hydrophobic neutrals) on XAD-8 resin for both Sides A and B of Lake Skjervatjern show no indication of initial breakthrough of these solutes. Therefore, the actual k'0.5 values for these hydrophobic solutes are greater than 50. There appears to be no losses of these solutes from the XAD-8 resins which would have been potentially sorbed onto the second resin, XAD-4. For this reason, the composition of the XAD-4 acids will not be influenced by varying amounts of hydrophobic constituents mixing into the XAD-4 acids, due to either exceeding k' ffi 50, or the small amount of hydrophobic constituents bleeding through the

Removal of organic solutes from water by resins

605

o. I c;m E

E:

qu,L~ 0

I

I

I

l

I

I

J

10 20 3o 40 5o 60 70 Column pore volumes of water passed through the XAD-8resin column

o

Fig. 5 (Part A). Idealized breakthrough curve for stream humic substances on XAD-8 resin (sample DOC of 10 mg C/L, hydrophilic DOC of 4 mg C/L, and humic substances are the primary constituent of the hydrophobic DOC). V= is the sample volume pumped through the XAD-8 resin column for a K'o.s = 50. At point V=, sample introduction ceases and back elution with dilute base would start.

~Or

,

~

Area representing sorbed hydrophobic solutes~ I / c: 51I

/

/

/

~

j J j / /

/

Area representing non-sorvea

4ILl

10

20

,

veil

J j j j / / ~ ~ Area representi ng / ~s'or~'ea'~hydrophobic A . ~ - ~ ~%1 ~S01utes between ki0.01

/

y

p

0.5

o O

~

30

40

50

I

1

60

70

Column pore volumes o f water sample passed through the XAD-8 column Fig. 5 (Part B). Idealized breakthrough curve for stream humic substances on ADX-8 resin (sample DOC of l 0 mg/L, hydrophilic DOC of 4 mg/L, and humic substances are the primary c o n s t i t u e n t s of the hydrophobic DOC). V= is the sample volume pumped through the XAD-8 resin column for a k'e.s = 50. At V=, sample introduction ceases and back elution with dilute base would start. A is the point of first detectable solute breakthrough, k',.,1 = 50.

XAD-8 resin between initial breakthrough (k'0.0t = 50) and k'0.5 = 50. The idealized breakthrough curve for a given solute system such as fulvic acid is shown in Fig. 5, Part A (Malcolm 1991). At a k'0.5 = 50 (point VE on the break-through curve), approximately 5% of the total given solute in the sample has broken through from point A (initial detectable breakthrough of k'o.01 = 50) and is represented in the hashed area of Fig. 5, Part B. Without actual DOC breakthrough curves, or

other breakthrough curves, it is impossible to state that a given k' has been attained or exceeded. The k' of fulvic acids, humic acids, and hydrophobic neutrals has been found to be greater than 50 in every instance determined. There are possible exceptions, but a general k' of 50 for these solutes is well established to remove 95% or more of these respective solutes. Therefore, if the actual k' for fulvic acids, humic acids, and hydrophobic neutrals were riO, the breakthrough curve for each solute would be as the

606

R.L. Malcolm and P. MacCarthy

10

Area representing sorbed hydrohpobic solutes / v E ~ Area representing /--, / ~/_., . __.I ~. , j/"J ~ non-sorbed hydrophilic / / / J J / ~ / / / J / /~.. i solutes in excess I

E o

VE ' m

5

[ \ \ _\ \ \ \ \ \ \ \ \ \ \ \ I ,.'.".,,.. ,.. v , . . I [Area representing non-sorbed hydrophilic \ ~1 ./ , i .~ .1 ~-" ,," j ' |. \ \ . . \ solutes atnormal k'^ ~ = 50 xl ~- ~ v .I 1 1 - - I

e--

4i,i

, \ , \ , \ _ \\

J,,. 0

" "

0

l

" I

\ \ I

".X

., I\ \

" I ~

~

I

~

~I

~0 20 ao 40 50 60 70 Column pore volumes o f w a t e r sample passed through t h e XAD-8 r e s i n column

Fig. 5 (Part C). Idealized breakthrough curve for stream humic substances on ADX-8 resin (sample DOC of 10 mg/L, hydrophilic D o e of 4 mg/L, and humic substances are the primary constituents of the hydrophobic DOC). Vs is the sample volume pumped through the XAD-8 resin column for a k',.s = 50. V,' is the actual sample volume passed through the XAD-8 column. At Vs', k',o = 25.

idealized k'o.5 = 50 breakthrough curve and 5% of the fulvic acids, humic acids, and hydrophobic neutrals would be lost from the XAD-8 column before sample introduction had stopped, and 5% of each of the solutes would be available for sorption onto the second resin column, XAD-4. This would result in the possibility that the sorbed isolate of XAD-4 acids eluted by dilute base would include a large amount of the low-molecular-weight acids which were hydrophilic or unsorbed on XAD-8 and co-mingled with the 5% of both fulvic and humic acids lost from the XAD-8 column. If any or all of the hydrophobic solutes of fulvic acids, humic acids, or hydrophobic neutrals from any given sample had a k' less that 50, then the XAD-4 isolate would be co-mingled with a greater amount of the lower k' solutes as shown in Fig. 5, Part B. A much more common and probable situation is shown in Fig. 5, Part C. In this instance, a k' of 50 was ignored, too much sample was pumped into the XAD-8 column, and a considerable amount of the hydrophobic isolates of k' = 50 was leached from the XAD-8 resin and passed onto the XAD-4 resin for possible sorption. If the XAD-8 resin was the only resin used for the isolation procedure, the hydrophobic solutes retained on the XAD-8 would have a k' of greater than 50, and there would have possibly been a fractionation of fulvic acids, humic acid, and hydrophobic neutral solutes with loss of lower k'

fractions, and the hydrophobic isolates would be attained but they would be potentially different in composition and character than a k' = 50 isolate. If XAD-4 was used following the XAD-8 resin, the XAD-4 isolate composition would be variable, dependent on to what extent the k' of 50 was exceeded on the XAD-8 resin. In other words, direct comparisons of XAD-8 and XAD-4 isolates from different sources would be invalid without proper quality control in the use o f the resins. CONCLUSIONS

The combined XAD-8 and XAD-4 resin procedure for the isolation of dissolved organic solutes from water was found to isolate 85% or more of the organic solutes from Lake Skjervatjern in Norway. Approximately 65% of the DOC was first removed on XAD-8 resin, and then an additional 20% of the DOC was removed on XAD-4 resin. Approximately 15% of the DOC solutes (primarily hydrophilie neutrals) were not sorbed or concentrated by the procedure. Of the 65% of the solutes removed on XAD-8 resin, 40% were fulvic acids, 16% were humic acids, and 9% were hydrophobic neutrals. Approximately 20% of the hydrophilic solutes that pass through the XAD-8 resin were sorbed solutes on the second resin, XAD-4 (i.e., they were hydrophobic relative to the XAD-4 resin). The fraction sorbed on XAD-4 resin was called XAD-4 acids because it

Removal of organic solutes from water by resins

represented approximately 85-90% of the hydrophilic XAD-8 acid fraction according to the original XAD8 fractionation procedure. The recovery of hydrophobic acids (fulvic acids and humic acids) and the hydrophobic neutral fraction from XAD-8 resin was essentially quantitative at 96%, 98%, and 86%, respectively. The recovery of XAD-4 acids from the XAD-4 resin was only about 50%. The exact reason for this moderately low recovery is unknown, but could result from r~-~ bonding between these organic solutes and the aromatic matrix of XAD-4. Precise and accurate methods of DOC determination must be available to the researcher for analytical as well as geochemical reasons. Precise DOC values to establish natural DOC concentrations, to enable accurate DOC effluent concentrations for DOC breakthrough curves, and to define other geochemical parameters are essential for meaningful geochemical data inperpretations. Without precise DOC data, hydrophobic and hydrophilic data should be considered only as approximate. From a geochemical perspective, although Side A and Side B of Lake Skjervatjern exhibited a 20% variation in total DOC concentration, the respective fractions upon isolation were almost identical in relative percentages of the total. The hydrophobic/ hydrophilic separation on XAD-8 resin for water from each side was also almost identical at 65% and 67%. These data support the contention that both sides of the lake are similar in general organic chemical composition. The precision of the sequential resin procedure has not yet been established, but it appears to be good, assuming that the organic constituents on both sides of the lake are similar.

607

REFERENCES Aiken, G.R.; Thurman, E.M.; Malcolm, R.L.; Walton, H.F. Comparison of XAD macroporous resins for the concentration of fulvic acid from aqueous solution. Anal. Chem. 51: 1799-1803; 1979. Gjessing, E.T. The HUMEX Project: Experimental acidification of a catchment and its humic lake. Environ. Int. 18: 535-543; 1992. Leenheer, LA. Comprehensive approach to preparative isolation and fractionation of dissolved organic carbon from natural waters and wastewaters. Environ. Sci. Tech. 15: 578-587; 1981. Leenheer, J.A.; Noyes, T.I. A filtration and column adsorbent system for on-site concentration and fractionation of organic substances from large volumes of water. U.S. Geological Survey, Water Supply, Paper no. 2230. Washington, DC: U.S. Government Printing Office; 1984. Malcolm, R.L. Factors to be considered in the isolation and characterization of aquatic humic substances. In: Boron, H.; Allard, B. eds. Humic substances in the aquatic and terrestrial environmont. London: Wiley; 1991: 369-391. Malcolm, R.L. The geochemistry of stream fulvic and humic acids. In: Aiken, G.R.; McKnight, D.M.; Wershaw, R.L.; MacCarthy, P., eds. Humic substances in soil, sediment, and water: geochemistry, isolation,and characterization. N e w York., NY: Wiley; 1985; Chap. 7: 181-209. Malcolm, R.L.; Thurman, E.M.; Aiken, G.R. The concentration and fractionation of trace organic solutes from natural and polluted waters using XAD-8, a methylmelhacrylate resin. 1 Ith annual conference on trace substances in environmental health. Columbia, M O : University of Columbia; 1977: 307-314. Mantoura, R.F.C.; Riley, J.P.The analyticalconcentration of humic substances from natural waters. Anal. China. Acta. 76: 97-106; 1975. Packham, R.F. Studies of organic colour in natural water. Prec. Soc. Water Treat. Exam. 13: 316-334; 1964. Thurman, E.M.; Aiken, G.R.; Malcolm, R.L. The use of macroreticulate resins to preconcentrate trace organic acids from water. Prec. 4th joint conference on the sensing of environmental pollutants, N e w Orleans, LA, Nov. 1977; Paper no. 166: 630-634. Available from: American Chemical Society, Washington, DC. Thurman, E.M.; Malcolm, R.L. Preparative isolation of aquatic humic substances. Environ. Sci. Tech. 15: 463-466; 1981, Thurman, E.M.; Malcolm, R.L.; Aiken, G.R. Prediction of capacity factors for aqueous organic solutes on a porous acrylic resin. Anal. Chem. 50: 775-779; 1978.