Organic desorption from carbon—I

Water Research Pergamon Press 1971. Vol. 5, pp. 3-18. Printed in Great Britain

ORGANIC DESORPTION FROM CARBON--I A CRITICAL LOOK AT DESORPTION OF U N K N O W N ORGANIC MATERIALS F R O M CARBON S. C. ALt.~, R. H. PAHL and K. G. MAVHAN Contribution No. 60, Graduate Center for Materials Research, Universityof Missouri-Rolla, Rolla, Missouri 65401, U.S.A. (Received 7 October 1970)

INTRODUCTION Dtnur~G work on the development of a method for the identification of microorganic impurities in D 2 0 process feedwater at the AEC Savannah River Project operated by E. I. DnPont for the Atomic Energy Commission, samples were collected by passing water through a carbon adsorption unit. Attempts to desorb the impurities from the carbon utilizing reported methods were not satisfactory. As a result, the parameters affecting the desorption of organics from carbon by means of extraction techniques were investigated. The bulk of this paper is concerned with the effects of residual moisture in carbon on chloroform and absolute methanol extraction efficiency. Drying methods and other parameters which affect the desorption process are also presented. STATE OF THE ART Numerous techniques have been developed for the determination of very low concentrations of organic compounds in water. These techniques each consist of a means of concentrating the organic substances, followed by instrumental analyses which include gas-liquid chromatography (BAKe, 1965, 1966, 1967; C~atUSO, 1966; RYCICMAN, 1967), infra-red spectroscopy (MXDDLETON, 1956; ROSEN 1956), ultraviolet spectroscopy (Dom~mUSH, 1963), and mass spectrometry (I-L~P, 1957). The major problem has been to increase the concentration of the organic compounds present to levels detectable by the instrumentation employed, without destroying or altering their chemical composition. Proposed methods of concentration include adsorption on activated carbon followed by extraction with an organic solvent (AME~C~,~ PUBLIC HEALTH ASSOCIATION, 1965), liquid-liquid extraction ( C ~ u s o , 1966), freeze concentration (BAKER, 1965), and vacuum rotary evaporation (SLONIM, 1966, 1967). The most common methods employed for the concentration of organics, utilized by public health, minicipal, and commercial laboratories, are the Carbon Chloroform Extract (CCE) Method ( A ~ a c x N PUBLIC H ~ m ' H AsSocxA~ON, 1965) and the Carbon Alcohol Extract (CAE) Method. Before extraction with either solvent, the carbon must be dried. A review of the literature from 1950 to present revealed no basic studies concerning residual moisture effects on extraction efficiency. BRAYS et al. (1951) proposed an analytical method for the recovery of organics by carbon extraction with ether in 3

4

S.C. ALLEN,R. H. PAHLand K. G. MAYHAN

which the carbon was air dried on a 60-mesh bronze screen. The temperature, drying time, and bed height were not mentioned. MIDDLETONet al. (1952) thinly spread the carbon on a screen or plate and dried at room temperature until "the carbon appeared dry and flowed readily". They state the carbon may contain 25-40 per cent residual moisture, but also state, without verification, that this moisture has no effect on the extraction with chloroform. The methods employed and criteria for the degree of drying prior to extraction as specified in Standard Methods (AMERICANPUBLIC HEALTH ASSOCIATION,1965) are ambiguous. The method states in part, Spread the carbon out in a thin layer on a tray of impervious material such as copper or glass. Dry the carbon in clean surroundings. To speed drying, pass heated air (35°-40 C) over the trays. Avoid high heat. Regard the carbon as dry when it is free flowing and appears like fresh, unused carbon. This is a standard procedure for the determination of chloroform extractable materials per liter of filtered water for which the maximum limit is 200/~g 1-1 for drinking waters. Yet no effort has been expended to clearly specify a drying procedure including temperature, bed thickness, and drying time. Because of the lack of a specified drying procedure, there is a range of reported drying times from overnight (RosEN, 1956) tO 5 days (GRIGOROPOLrLOS, 1968). ROBINSON et aL (1967) placed the carbon in 2-in. thick layers on trays and dried two to three days in a constant temperature room at 95°F. The longest time for drying was reported by GRIGOROPOOLOSand SMITH (1968). The carbon was dried in wooden trays lined with polyethylene for five days at 40°C. The bed depth, however, was not reported. EXPERIMENTAL PROCEDURES

Sampling The procedure utilized to collect samples was essentially the same as that described in Standard Methods (AMERICANPUBLIC HEALTHASSOCIATION,1965). A 4-in., schedule 40 stainless steel pipe 4 ft in length was installed in the D 2 0 process feedwater line as shown in FIG. 1. The water was taken from the Savannah River and treated by coagulation, filtration, deaeration and pH adjustment, prior to processing. The adsorption cylinder unit was filled with 3.5 ft (0.30 ft a) of carbon (Pittsburgh Carbon Type CAL 12 × 40 mesh). The inlet water pressure was maintained at 40 psig with an average flow rate of 0.31 gpm. After about 14,000 gal of water had passed through the carbon, the adsorption unit was removed from the system. The unit was drained of free flowing water and shipped to UMR* for analyses. The carbon was removed from the cylinder and stored in a sealed heavy polyethylene bag prior to testing.

Drying techniques Three methods of carbon drying were investigated. Drying curves were obtained for (a) freeze-drying, (b) low pressure rotary-flash evaporation and (c) air drying. In any given procedure the sample weight, size and bulk configuration were maintained constant. Samples used in the different procedures were kept as similar as possible for comparative purposes. * University of Missouri-Rolla

Organic Desorption from Carbon--I

SA.

5

To

COLLECTOR FLANGED JOINT PRESSURE GAUGE

~

~ WATER NEEDLE

DRAINVALVE

VALVE" ~ DRAN IVALVE 1:I(3. l. Sample coUcction schematic.

(a) A VirTis Freeze-Mobile was used in all of the freeze drying trials. This unit has 12 one-half-inch ports and operates in the 10- a mm Hg range at coil temperatures to --55°C. In all trials, 600 ml drying flasks, 90 mm i.d., were used and in the tests reported, only one flask was processed at a time to insure continuity of results. Wet carbon (130 g) was charged into the flask, and the flask was immersed in liquid nitrogen and allowed to remain immersed for 10 min until the water portion of the carbon mix was frozen. A l~riod of I0 min allowed the center of the carbon to reach a temperature of --60°C. The flask was then connected to one of the ports on the freeze drier head. Drying time was then recorded from the point when vacuum was applied to the flask. Weighings were made as a function of time until a constant weight was obtained. The initial drying curves were obtained on wetted carbon blanks, and subsequent samples containing the adsorbed material showed the same drying characteristics. All water removed from the carbon was condensed and collected for further analysis in the event of removal of high vapor pressure constituents from the carbon. In the cases of the carbon blanks, a thermogravimetric trace was made on the freeze dried samples to determine the amount of bound moisture remaining. It was found that less than 2 per cent moisture remained which could not be removed until the sample temperature approached the normal boiling point of water. Subsequent thermogravimetric traces of carbon containing adsorbed materials indicated that much of these materials would be lost in attempting to remove the last traces of water. It should be emphasized that the constant weight lines on the drying curves do not imply complete removal of water from the carbon. A compilation of points for several runs are presented in FIG. 2. (b) An all glass-teflon constructed, low pressure rotary and flash evaporator (Chomquip Company) was also utilized to dry the charcoal. The unit was equipped with

6

S . C . ALLEN, R. H. PAHL and K. G. MAYHAN

Got

x

13o

~40

t 20 i

.ol/ 0

V-

I

I

I

4

I

I

I

I

I

I

I

8 12 16 20 FREEZEDRYINGTIME- HOURS

I

24

28

Fie. 2. Freeze drying curve.

a Pyrex cold finger condenser 400 mm in length with a one liter receiver; the receiver was fitted with a teflon stopcock to facilitate removal of the condensed vapors. There are several advantages to using this type of evaporator. Evaporations can be carried out in a batchwisc manner or by continuous addition. More important, in this case, all of the evaporated material can be condensed and recovered for analysis to determine what has been rcmov~l from the sample. It was found that the final moisture content of the carbon could not be reduced quite as much with the evaporator as could be obtained with the freeze drier but that the initial moisture content could b¢ reduced much more rapidly using the evaporator. In order to conserve time, a sample can be initially evaporated and then transferred to the freeze drier until a constant weight is obtained. A characteristic vacuum evaporator plot is presented in FIG. 3. (c) Samples were also air dried in one inch depths, utilizing polyethylene dishes in a natural convection oven, at 35°C. Weight loss was recorded as a function of time. The

I00

5O

8O

xx

4O

j~ ss S

•

W a,.

i s/

"~x

s0

•

o I

3O

/

..r -q

xx s S~ xJ

,.J

xx

/"

W

zo g

%%

' f

f s

"%,

I0

~" 2O

"st i__

"

.

.

2.

.

.

4.

6

'

'

)'o';2

DRYINGTIME,ROTARYEVAPORATOR-HOURS FIG. 3. Rotary evaporation drying curve.

14

Organic Desorption from Carbon--I

7

one inch depth of carbon was entirely arbitrary, and different results can be expected using other sample depths. A series of tests was set up to demonstrate the degree of drying that can be obtained with various bed thicknesses as a function of time. In order to illustrate the drying effects on layers of a given bed thickness, a series of Standard Tyler screens was used to contain multiples of 1 in. layers of carbon. FIOtr~ 4 shows the arrangement of the sieves and the bottom pan. With this arrangement it was possible to weigh the individual trays and thus determine weight loss as a

TYLER - STANDARD SCREENS

K SEALING TAPE

t FIG. 4. Simulated air drying apparatus.

function of time. During any given drying cycle the contact points between the sieves and the atmosphere were sealed such that diffusion of moisture from the charcoal had to be through the top layer of carbon. The distance between the carbon layers on adjacent screens was minimal (barely touching initially); this setup closely approximated a continuous carbon layer. F]Gur.Es 5-8 show air drying characteristic curves for 1 in., 2 in., 3 in. and 4 in. depth carbon beds, respectively.

Extraction procedure Extractions were carried out in the same 50-mm dia. Soxhlet extractor equipped with a Friedrichs condenser and a 500-ml round-bottom boiling flask. A 4-ram teflon stopcock was attached to the boiling flask, as shown in Fig. 9, to permit sample removal for u.v. analysis during extraction. A 45-ram dia. fritted glass thimble (very coarse) was used to hold the carbon in the extractor. The flask was heated with a 50

4O

cn 30

g .J

zo ,o

i 0

~) 2

i '

~

'

0

~iO

*

' t I I 80 I00 AIR DRYING T I M E - H O U R S

I 120

FiG. 5. Air drying characteristics, 1 in. bed, 35°C.

I 140

8

S . C . ALLEN, R,. H. PAHL a n d K. G. MAYAAN 50 .~

2" LEVEL

0

m 3=10

J

0

20

40

60

80

I00

120

140

160

AIR .DRYING T I M E - HOURS

FIG. 6. Air drying characteristics, 2 in. bed, 35°C.

40

3O U~ 0

t_ 20 -r u,I

---~'o' . ' ~

-.o'~s' ,~o ',~o"

FIG. 7. Air drying characteristics, 3 in. bed, 35°C. 50

40

3C o~ o~ O I-"lrc.9 LU

~10

0

20

40

60

80

AIR DRYING T I M E -

I00 HOURS

120

PIG. 8. Air drying characteristics, 4 in. bed, 35°C.

I

140

IG

Organic Dcsorption from Cmrbon---I

9

mantle connected to a variable transformer. The vapor tube of the extractor was insulated with 0.25-in. foam rubber to prevent condensation of the solvent vapor. The solvent was agitated with a 2-in teflon-coated magnetic stirring egg and a variable speed magnetic stirrer. An initial solvent volume of 360 ml was used in all extractions. This volume assured that the liquid level never dropped below the top of the mantle when the extraction WATER IN

"--~

OUT - , - - - c

SOXHLET EXTRACTION APPARATUS

FOAM INSULATION

SAMPLE SYRINGE FLASK THERMOOOUPLE TO R E C O R D E R ~ ~ I t ~ - P

f

MANTLE

STIRRING

10 i,

w

STIRRER v :

Fxo. 9. Extraction apparatus.

thimble was at maximum capacity. A carbon volume of 110 ml assured the solvent level in the extractor covered the carbon just before syphoning. The syphon rate and pot temperature were determined as a function of variable transformer setting. An iron-constantan thermocouple, referenced to 0°C in an icewater mixture, was taped outside to the bottom of the boiling flask and connected to a recorder, which permitted a permanent record of pot temperature and frequency and number of cycles. The effects of stirrer speed, and substitution of boiling chips, on the flask temperature were also investigated. All solvents used in this investigation were distilled, gave a clean chromatograph pattern and were considered to be spectral grade.

10

S.C. AJ.t.E% R. H. PAHL and K. G. MAYHAN

Analyses After arbitrary syphon cycles during each extraction, a 1 ml sample was taken from the boiling flask by inserting the syringe needle through the teflon stopcock. The samples were injected into a clean 1-mm u.v.-grade quartz cell. All samples were analyzed in the u.v. region (400 m/L to solvent cutoff3 in double beam mode, with pure solvent as the reference, using a Perkin-Elmer Model 450 spectrophotometer. Before each sample analysis, the instrument zero and full scale attenuations were checked at a slit width of 0.26 mm. The filled sample and reference cells were inserted, and the pen was adjusted to 100 per cent transmittance at 400 m/z. The pen adjustment was made in order that a given extraction analysis could be normalized around a common starting point. A great number of separate lines are obviously obtained during an extraction which would be complicated to present in this paper. Instead, the transmittance was recorded at a point before solvent cut-off and correlated with the number of cycles and degree of drying for a given solvent. Typical curves for chloroform and methanol extraction are shown in FIGS. I0 and 11. In order to obtain the total amount of organic materials recovered during each extraction, the final volume of solution was measured and the weight was calculated by multiplying by the solvent density. The ratio, mg residue/g solvent, was determined by evaporating to constant weight a known initial weight of solution at ambient temperature in an aluminium dish. Knowing the dry volume of carbon extracted, the carbon filter volume, and the volume of water passed through the filter, the value

9o~l, ao

)

~

o - 1 2 HOUR FREEZE DRY

60

50

40

:50

20

I0 I

0

I

I0

!

l

20

v

I

1

I

I

I

=

t

I

l

=

I

30 40 50 60 70 80 CYCLES~ SOXHLET EXTRACTOR, CHLOROFORM

FIG. 10. Transmission vs. number of cycles, chloroform.

I

90

Organic Desorption from Carbon--I

11

,oo 8oll

0 - 12 HOUR FREEZE DRY • - 16

70

"

•-21 • - 24

"

z60 n. o

"I,5o O

~40 I-

u)

~20 n.,, I-

I0

I0

20

30 40 50 60 70 80 C Y C L E S - - SOXHL[T EXTRACTOR, METHANOL

FIG. ] 1. Transmission

w.

number of cycles, Methanol.

o 600 :I:

m

.J (n

--~ 4 0 0 Z ¢D

c-

..-c----------

O300

0 W

0 - CHLOROFORM e-.METHANOL

E 2oo X W

.='loo

al

i

I~.

!

,~,

FREEZE

o

I;

!

ao n

!

~a

n

a~

DRYING T I M E - - H O U R S

F~C;. ]2. Quantity of ~ c t e , d organics 'vs. freeze ( L r ~

time.

90

12

S.C. ALLEN,R. H. PAre. and K. G. MAYHAN I00 _.9.

¢tJ Z
~ 9o X W

• - METHANOL

W

-1 W

o

80

I

12

f

I

I

I

I

I

I

I

I

t

t

14 16 18 20 22 FREEZE DRYING TIME--HOURS

24

FIG. 13. Extraction efficiencyvs. freeze drying time. microgram extracted organics 1-1 water, was calculated. These data are presented graphically in Fins. 12 and 13 for freeze dried samples extracted with chloroform and methanol. Data for air dried samples are in TABLES 3 and 4 in the next section. DISCUSSION OF RESULTS When allowed to work with known materials, desorption phenomena of organicmaterials from carbon is difficult, at best, to quantitize because of the large number of parameters which can affect the experimentation. In most instances where water has acted as a carrier during the adsorption process, an attempt is usually made to remove as much water as possible prior to extraction. One must be concerned, where even semiquantitative results are desired, that the material adsorbed is not lost during the drying process. For this reason low temperature air drying has become an accepted procedure. Depending upon the drying temperature, time, air flow rate, and bed depth, carbon can retain copious quantities of water and yet appear dry to the touch, as evidenced by FIGs. 5-8. Even with the freeze drying and vacuum evaporation methods, reported here, it was not possible to remove all o f the water from the carbon samples. The results presented in FIGS. 2 and 3 show that freeze drying and rotary evaporation methods are capable of removing more water from the carbon in a given period of time than low temperature air drying and that the final moisture content of the carbon is lowest with the freeze drying technique. In the particular samples being processed, there was no evidence of loss of anything but water during the various drying procedures. However, consideration should always be given this problem prior to solvent extracti on. No attempt has been made in this paper to delve into solution theory and why one solvent should be better than another for a specific extraction. This work has led to a fundamental study of extraction (desorption) of known organic materials from carbon as a function of the molecular and thermodynamic properties of solvents and solutes, which will be reported as Part II of this series. In an effort to stay within the realm of accepted procedures, the preliminary extractions were carried out using chloroform (essentially water immiscible) and absolute methanol (water miscible). By taking

Organic Dosorption from Carbon---I

13

advantage of the various drying procedures available it was possible to obtain extraction data on samples containing the material to be extracted along with different quantities of water. During the extractions the maximum temperature of the pot was of interest. The bulk temperature of the boiling solvent is essentially constant at a given pressure but the pot temperature and rate of solvent vaporization is dependent upon heat input to the system. Thus, for a constant boiling liquid the temperature at the pot wall can vary over wide ranges. In those instances where decomposition of the extracted material could be a problem, temperature control should dictate the other extraction variables. TABLE 1 shows the effect of voltage input on pot temperature and cycle time at a constant stirring speed. These data show that temperatures at the pot wall can be almost 200°C above the normal boiling point of the solvent and that an arbitrary power input to the heating mantle could be undesirable for some extractions. TABLE2 shows indirectly the effect of stirrer speed on the pot temperature. With the particular magnetic stirrer used in these experiments it was not possible to obtain speed in revolutions per unit time (the stirrer housing was a rivited assembly and a response could not be observed by the strobe light through the solution), thus, only relative stirrer TABLE 1. VARIAC SETTING VS, ]POTTEMPERATUREAND CYCLE TIME

Solvent Chloroform normal boiling point, 61,2°C, AHo _-- 59 cal g -1

Methanol, normal boiling point, 64.7°C, AHv = 284.3 cal g - i

Variac setting

Pot temp (°C)

Cycle time (min)

Stirrer set

40

117

29

6

50

148

16

6

60

182

10

6

70

218

7.5

6

80

258

5.5

6

50

135

31.5

6

60

166

20

6

70

195

14.5

6

80

233

11

6

TABLE 2, EFFECT OF STIgRER SPEED ON POT TEMPERATURES Solvent

Variac setting

Stirrer set

Pot t=-mp (°C)

Chloroform Chloroform Chloroform

48 48 48

3 6 8

142 138 136

Methanol Methanol Methanol

60 60 60

3 6 8

179 177 176

12 16 21 24

Methanol Methanol Methanol Methanol

* 104 cycles

12 16 21 24

Chloroform Chloroform Chloroform Chloroform

Solvent

Freeze dry time (h)

28 14 6 3

23 12 7 3

Residual moisture (%)

0.098 0.085 0.070 0.062

T after 80 cycles at 220 mr,

0.341" 0.350 0.340 0.290

T after 80 cycles at 255 mt~

1.009 1.071 1.155 1.208

0.467 0.456 0.469 0.538

Absorbance

TABLE 3

1.45 1.52 1.67 1.76

0.52 0.55 0.57 0.67

mg Organic g Solvent

328 331 341 364

196 210 224 234

Total (nag)

491 495 510 545

293 314 335 350

t~g organic 1. HsO

0.696 0.706 0.691 0.689

Z

"<

0

go

.~ 0.804

~v

.~

0.834 0.824

--

l) 0.903

Apparent extinction cocf. (g m g - i m m -

X

Organic Desorption from Carbon--I

15

settings can be reported. Several other runs were made where the magnetic stirrer was replaced with boiling chips and pot temperature and cycle times were recorded. It is interesting to note that the pot temperature and cycle times were only slightly higher using the boiling chips. FIGUP.ES 10 and 11 show the progress of extraction in chloroform and methanol as detected by the spectrophotometer. For the chloroform runs, a stirrer setting of 6, a variac setting of 48, and a cycle time of 16--17 min was utilized; for the methanol runs, a stirrer setting of 6, a variac setting of 60, and a cycle time of 21 rain was utilized. The data show that in both cases the greater the amount of residual moisture in the carbon the more retarded the extraction. FlOUZr.s 12 and 13 indicate that extraction with the water miscible methanol extracts more material than with the water immiscible chloroform. Since the material in question is composed of unknowns, no statement can be made at this point concerning extraction efficiency or even whether the same materials are being extracted. TA~SL~ 3 gives a compilation of the chloroform and methanol extractions as a function of freeze drying time and per cent residual moisture. The transmittance after 80 extraction cycles was read at 255 mtz for the chloroform extracts and at 220 mtz for the methanol extracts. The only exception was the 12-h freeze dried chlorofolm extraction which was continued for 104 cycles. The transmittance was converted to absorbance using Lambert's Law: A ---- loglo (l/T) where: A ---- absorbance T ---- transmittance The residue concentration in the solvent and total residue extracted were calculated using the method described earlier. The ratio, micrograms of extracted materials per liter of filtered water, was calculated using the following unit equation: ~g = (mgexlacted)(-~g)(1.--HzO)(Vol. Filter Volume 1.H20 Extracted Carbon] These values are plotted as a function of freeze drying time for both solvents in FIG. 12. Assuming a relative efficiency for the 24-h freeze dried carbon extraction, the per cent materials extracted are plotted in FIG. 13. The chloroform extraction efficiency shows a linear increase of 18 per cent with freeze drying time. The methanol extraction efficiency increases by only 10 per cent, but there is a very sharp upward trend after 21 h of freeze drying. An apparent extinction coefficient was calculated using Beer's Law: A = abe

where: A ~- absorbance a ---- path length, 1 ram b ---- extinction coefficient, g solvent/mg res-mm e ---- concentration, mg residue g-1 solvent

12 16 21 24 115

12 16 21 24 115

* Calculated after 80 cycles.

Methanol Methanol Methanol Methanol Methanol

Solvent

Dry time (h)

* Calculated after 104 cycles. t Calculated after 80 cycles.

Chloroform Chloroform Chloroform Chloroform Chloroform

Solvent

Dry time (h)

28.2 14.1 6.2 3.1 7.2

(%)

Residual moisture

23.2 12.2 7.5 2.8 7.2

(%)

Residual moisture

0.375 0.338 0.308 0.290 0.327

T After 12 cycles at 220 mr,

0.670 0.623 0.600 0.545 0.625

Tafter 12 cycles at 255m~

0.426 0.471 0.511 0.538 0.485

TABLE 5

0.174 0.206 0.222 0.264 0.204

TABLE 4

0.426 0.471 0.511 0.538 0.485

bc ( m m - ~)

0.174 0.206 0.222 0.264 0.204

be (ram- ~)

i mm-l) 0.696* 0.706* 0.691" 0.689* 0.688

b(gmg

0.748

0.8o4t

0.903* 0.834t 0.824t

Apparent extinction coef. (g rng -~ m m -1)

0.613 0.667 0.740 0.780 0.7046

mg Organic g Solvent

0.193 0.247 0.269 0.328 0.2726

mg Organic g Solvent

,<

P

t-

z

I-'

>

p

F~

Organic Desorption from Carbon--I

17

In order to extract qualitative information from this work it was necessary to take the liberty of assuming that the calculated extinction coefficient was a constant for a given extraction. TAnL~S4 and 5 compare the freeze dried carbon with 115-h air dried carbon at 35°C after extraction for 12 cycles with chloroform and 16 cycles with methanol. The transmittance was read at the same wave numbers as before. After converting the transmittance to absorbance, the bc product, ram-1, could be calculated as the path length was constant. Using the apparent extinction coefficient calculated after 80 cycles (for freeze dried carbon), the 12-cycle chloroform extract concentrations and 16-cycle methanol extract concentrations were calculated for comparison with the measured concentrations for the air dried carbon. The calculated concentrations and measured concentration for a given solvent extraction appear to be strongly dependent upon the percentage residual moisture of the carbon after drying. The apparent extinction coefficient appears to remain constant for a given extraction (regardless of the number of cycles) and is also a function of the residual moisture of the carbon. CONCLUSIONS

It has been shown conclusively that the amount of residual moisture remaining in carbon samples after drying is a function of the drying technique. In particular freeze drying > rotary evaporation ~ air drying for both rate and efficiency of drying. Drying procedures which have been adopted by USPHS (A~mUCAN PUBLIC ~ T H ASSOCIATION, 1965) can lead to a wide range of varied results. It would appear that a more precise drying procedure is required before reproducibility can be expected between investigators. The tentative Standard CCE Method (AMERICAN PUBLIC H~CLTH ASSOCIATION, 1965) is not specific from the standpoint that the operating procedure specified is subject to wide interpretation. As a consequence reproducibility of results between different investigators and laboratories may be fortuitous. The work presented in this paper shows that the amount of residual moisture present in the carbon is a major parameter which affects Soxhlet extraction df~iency. The power input put to the heating mantle is another parameter which could have an effect on the final extraction results since pot wall temperatures can be obtained which could lead to thermal degradation of the materials being extracted. It can be concluded that the ideas behind the tentative Standard CCE Method are sound but that the operating procedures within the Method for drying and extraction are not specific. Until some clarification of the accepted procedures are set forth, the 200 pg l - i limit for CCE extractables in drinking water results obtained may be in error when considered on an absolute basis. AcknowledgementmThis work was sponsored by the U.S. Atomic Energy Commission and the Graduate Center for Materials Research, University of Missouri-Rolla. The authors also wish to express their gratitude to the DuPont 1~rsonnel at the Savannah River Plant in Aiken, South Carolina for their assistance and suggestions.

REFERENCES AMERICANPtmLIC HEALTHASSOC,'noN (1965) Standard Methods for the Examination of Water and Waste Water, 12th edn, pp. 214--218. BAKERR. A. (1965) Microchemical contaminants by freeze concentration and gas chromatography. J. Wat. Pollut. Control Fedn 37 (8), 1164-1170. WATER5limb

18

S . C . ALLEN, R. H. PAHL and K. G. MAYHAN

BAKER R. A. (1966) Trace organic analyses by aqueous gas-liquid chromatography. Int. f. Air Wat. Pollut. 10, 591-602. BAmm R. A. and MALO B. A. (1967) Characterization of the microorganic constituents of water by instrumental procedures. Wat. Sewag. Wks R-223-230. BRAOSH., MIDDL~TONF. M. and WALTONG. (1951) Organic chemical compounds in raw and filtered surface waters. AnaL Chem. 23, (8), 1160-1164. CARUSOS. C., Bg,o~rt H. C. and HOAK R. D. (1966) Tracing organic compounds in surface streams. Int. J. Air. Wat. Pollut. 10, 41-48. CrmMQtrm COM'PANY,Berkeley, California, Model 100. Do~u,musH J. N. and RYCKMAND. W. (1963) The effects of physiochemical processes in removing organic contaminants. J. Wat. Pollut. Control Fedn 35 (10), 1325-1338. GmooRoPoOLos S. G. and SMrm J. W. (1968) Trace organics in Missouri subsurface waters. J. Am. Iv'at. Wks Ass. 60, 586-596. I4_~I, G. P., S'rEW~T D. W. and COOPER H. C. (1957) "Mass Spectrometric Determination of Volatile Solvents in Industrial Waste Water." Anal. Chem. 29 (1), 68-71. MIDm.mrON F. M., Bl~,uS H. and R u c n n o r r C. C. (1952) The application of the carbon filter and eountereurrent extraction to the analysis of organic industrial wastes. Proe. Seventh Purdue Univ. Ind. Waste Con. 79, 439--454. MIDDL~rON F. M., GRANT W. and ROSEN A. A. (1956) Drinking water taste and odor-corelation with organic chemical content. Ind. Engng Chem. 48, 268-274. ROmNSON L. R., JR., O'CONNOR J. T. and ENGELBRECHTR. S. (1967) Organic materials in Illinois ground waters, ar. Am. Wat. Wks Ass. 59, 227-236. RosE~ A. A., MIDm.m'ON F. M. and TAYLORN. W. (1956) Identification of anionic synthetic detergents in foams and surface waters..Jr. Am. Wat. Wks Ass. 48, 1321-1330. R Y C X a ~ D. W., IatWIN J. W. and Yotr~o R. H. F. (1967) Trace Organics in surface Waters. or. Wat. Pollut. Control Fedn 39 (3), 458-469. St.ON~ A. R. (1967) Analysis of total solids by the vacuum rotary evaporation method, or. Wat. Poilut. Control Fedn 39 (5), 846-849. SLO~m~ A. R. and C ~ w ~ Y F. F. (1966) Precise determination of totalsolids in water, or. Wat. Pollut. Control Ass. 38 (10), 1609-1613.