Orangic Chemical Compounds in Raw and Filtered Surface Waters

May 1, 2002 - Analytical Chemistry 1952 24 (12), 1872-1876 .... Journal of the Marine Biological Association of the United Kingdom 1952 31 (02), 335...
0 downloads 0 Views 651KB Size
A N A L Y T I C A L CHEMISTRY

1160 In order to estimate the diffusion rate of solvent through the connecting tube, the following experiment was carried out. One end of a 2-foot length of borosilicate glass capillary, 1.25 mm. in bore, was drawn out to a fine tip. The other end was fastened through a 50-ml. filter flask as shown in Figure 6. Hydrogen gas was passed through the apparatus and forced through the capillary. The fine tip was dipped in methanol and a column of solvent allowed to rise in the capillary until the meniscus was 20 mm. above the drawn-down section; then the tip was removed and quickly sealed in the Bunsen flame. I n this way, the solvent was sealed in the capillary under hydrogen. Finally, the capillary was mounted in a vertical position, and hydrogen was passed slowly (0.5 ml. per minute) through the filter flask for several days. The average rate at which the nieniscus receded was 2.5 mm. per day a t 25°C.; this rate was not changed measurably during an additional 24 hours with an increased hydrogen flow of 15 ml. per minute. In a control experiment, the capillary was evacuated, whereupon the methanol meniscus receded at a rate of 3 mm. per minute. The slow diffusion of methanol through hydrogen was thereby shown not to be a result of limitation by the extent of evaporation surface. These results suggest that in this case correction for solvent vapor pressure is not justified because no significant amount of vapor can enter the measuring system under the specified operating conditions. In the literature on laboratory hydrogenation at constant pressure, most experimenters who made a correction for solvent vapor had previously saturated their hydrogen before introducing it into their apparatus. The closest analogy to the authors’ apparatus is described by

Jackson ( I ) , who used a flexible glass helix and other connection6 totaling 3.4 meters of tubing 5 mm. in inside diameter (estimated) between measuring system and reactor. He did not saturate his hydrogen with solvent, but did make a correction for its vapor pressure. This correction may be justified because solvent vapor could reach his measuring system through a smaller diffusion barrier than the authors’, a t 1 atmosphere hydrogen pressure, during the 3 to 4 hours that Jackson alloxed for temperature equilibration. ACKNOW LEDGJIENT

The authors are grateful to Lawrence T. Hallett for his interest and encouragement in this work, and to J. P. G. Beiswanger, D. L. Fuller, and Clyde McKinley for many helpful conversations on the diffusion effects discussed above. To F. C. Snowden, and the draftsmen and machinists of this laboratory, the authors express their thanks for help in the design and construction of the apparatus. LITERATURE CITED

(1) Jackson, H., J . SOC.Chem. Ind.,57,97T (1938). (2) Johns, I. B., and Seiferle, E. J., IND.ENG. CHEM.,. ~ N A L .ED., 13, 341 (1941). (3) Joshel, L. M.,Ibid.,15, 590 (1943). (4) Ogg, C.L., and Cooper, F. J., rlwar. CHEY.,21, 1400 (1949). (5) Whitmore, W. F., and Revukas, A. J., J . Am. Chem. SOC.,59, 1500 (1937); 62,1687 (1940). (6) Zaugg, H.E.,and Lauer, W. M.,.IN.~L. CHEY.,20, 1022 (1948). RECEIVED Febiuary 10, 1951.

Organic Chemical Compounds in Raw and Filtered Surface Waters HARRY BRAUS, F. 31. MIDDLETON, AND GRAHAM WALTON U . S . Public Health Sercice, Cincinnati 2, Ohio KXORLEDGE of the kinds and concentrations of organic chemical compounds in surface waters is important in studies concerned with tastes and odors in drinking water, natural purification of streams, analysis and tracing of industrial wastes, and toxic and other physiological effects on man and animal. Direct determination of most organic compounds is not usually possible because of the minute concentrations that normally occur in surface waters. This study describes a method for the concentration and estimation of organic compounds in raw and filtered surface water. Possible applications of the techniques presented are discussed.

Figure 1. It is pieceded 1)san iroii cylinder approximately halffull of filter sand and fine gravel. Valves are arranged so that the sand filter may be by-passed, back-washed with the raw water, or rinsed by filtering to waste. 9 minimum head of 10 feet of water is necessary for the effective operation of this unit. A plastic filter, used for Inhorntory studies, consists of a 4foot

CONCENTRATION O F ORGANIC COMPOUNDS FROM RAW AND FILTERED SURFACE WATERS

The concentration of the organic compounds in the waters tested is accomplished by passing from 5000 to 75,000 gallons of water through small portable activated carbon filters. The average rate of flow through the filter is approximately 0.1 to 0.6 gallon per minute. Figure 1 shows the simplest type of activated carbon filter for use on filtered or finished water. This unit consists of a %foot length of iron pipe 4 inches in diameter. Between 1200 and 1500 grams of granular activated carbon are charged into this cylinder. The carbon is held in place by perforated cadmium-plated disks, equivalent to about a 16-mesh screen, soldered into the reducers a t each end. Figure 2 shows the assembly used when turbid or raw waters are sampled. The carbon filter is the same its that shown in

WATER METER F l N l S H E D WATER

I’

e!I

EFFLUENT - J

INFLUENT

Figure 1. Carbon Filter Installation Uhed for sampling filtered water for trace quantities of organic chemicals

V O L U M E 23, NO. 8, A U G U S T 1 9 5 1 This study was undertaken to develop a method for the recovery and identification of minute quantities of organic chemicals in drinking water and drinking water sources. Significant quantities of organic chemical compounds w-ere recovered by passing water through activated carbon filters, followed by drying of the carbon and elution with an organic solvent. The organic residue was separated into five groups and certain specific chemicals were identified. Threshold odor tests were run on selected

1161 samples. For the first time, the effects of minute quantities of organic chemicals in water supplies can be studied in relation to: (1) the causes and treatment of objectionable tastes and odors in water, (2) industrial waste contamination in streams, (3) pollution caused by the growth and decay of aquatic plants and animals, and (4) possible biological or toxic effects on man or animals. The technique may be applied directly to industrial wastes of certain types and to other materials in dilute solution.

Yuchar (2-190 carbon has less than one fourth as much ether extractable material as does the Cliffchar carbon. i t is not believed that the extracted materials interfere p i t h the test made on the total organic residue. A comparison of the adsorptive ability of these carbons was made using Cincinnati t a p 15ater. Approximately 8000 gallons of water were passed through each of two filters containing the carbons during the same period (Table 11). Approximately five times as much material was recovered from the same quantity of water using the C-190 carbon. STUDIES ON FIELD FILTERS

Filters of the type described were placed in n-ater plants a t various cities o n the Ohio River and a t other locations having surface water supUsed f o r s a m p l i n g t u r b i d w a t e r f o r t r a c e q u a n t i t i e s of o r g a n i c c h e m i r a l s plies. h effort \vas made to include certain locations known to be subject to industrial waste pollution and others known t o be free of such pollution. Relength of plastic tubing, 3 inches in diameter. Carbon is held in sults of these studips are discusscd helow. place by a 60-mesh bronze screen a t earh end. Figure 2.

Sand and Carbon Filter Installation

SELECTION OF CARBONS

\-arious types of activated carbon \\ere tested for use in these experiments. Cliffchar (4- to 10-mesh) granular carbon was used in the filters placed in the field. A more active carbon of finer mesh would have been desirable for these studies, but to maintain satisfactory flows the coarse carbon \!as used. Studies of finished w-ater under laboratory controlled conditions were made with unground C-190 activated carbon, furnished by the industrial Chemical Sales Division of \Vest \-irginia Pulp and Paper Co.

Table I.

Ether-Extractable Materials in Two Activated Carbons

Carbon Cliffchar, granular h-uchar C-190, unground

Ether-Extractable Materials, 70 0.0117 0.0027

Two laboratory filters, 15 x 3 inches, were set up in series, and 12,000 gallons of Cincinnati drinking water were passed through them a t a rate of 0.2 gallon per minute. A total of 3.25 grams or 70 parts per billion of material were extracted from the f i s t filter in the series. The second filter, which received the effluent from the first filter yielded 1.26 grams of extract or 28 parts per billion. These carbons contain some ether-extractable materials as received from the manufacturer. Results of ether-extraction tests made with carefully purified ether are shown in Table I.

Table 11. Organic JIaterial Extracted from Equal Volumes of Same Cincinnati Tap Water by Two Types of Activated Carbon Type of Carbon Nuchar (2-190 Cliffchar

Weight of Carbon Used Grams 902 1900

Total Organic Material Extracteda Grams P . p . b .b 4.016 133 0.81 27

weight after correction for organic extractable material from carbon. * PNet a r t s per billion.

AN 4LYTICA L PROCEDURES

hftcr a quantity of water has been passed through the filters, they are returned to the laboratory and dismantled. The carbon is removed, air-dried on a 60-mesh bronze screen, then placed in a large-capacity modified Soxhlet extractor (Corning S o . 3885, Catalog LP28), and extracted with diethyl ether. A good grade of ether is selected. Care must be used t o keep a t a minimum the formation of peroxides in the ether. The U.S.P. test for peroxides in ether should be applied. Purification of ether may be accomplished as follows: The ether is shaken with acidulated ferrous sulfate, washed with distilled water, and dried with anhydrous calcium chloride. After standing several days over the calcium chloride, the ether is filtered through Whatman S o . 12 filter paper and distilled. Great care must be exercised in starting the extraction, because a large amount of heat of adsorption is evolved. The authors have found it reasonably safe to add carbon slowly to the cold ether previously placed in the Soxhlet extractor. After extraction, the ether extract is evaporated to a volume of 25 ml. Anhydrous sodium sulfate is added to the ether solution and the whole is allowed to 4 3 hours. The ether

A N A I. Y T I C A L C H E M I S T R Y

1162

tion is now extracted with four 10-ml. portions of 5% sodium hydroxide containing 20% sodium chloride. A soapy emulsion formed a t this point may be broken by the addition of a drop of ethyl alcohol or a little more water. The ether layer is dried with anhydrous sodium sulfate, filtered, and evaporated. The residue remaining is weighed as a mixture of neutral organic compounds.

Table 111. Total Anhydrous Organic Extract in Various Surface Water Supplies Location of Water Plant

Filter in Service From TO 2/14 2/2 5/28 1/24

Midland, Pa. Wheeling, W. 1’8. Huntington, W. Va. Pomeroy. Ohio Cincinnati, Ohio Louisville, Ky. Columbus, Okio

%4 5/28

.7/14 8/9 7/25 5/11 8/ 1 8/3 9/15

Sampling Point

Quantity of waters

Filter effluent Filter e5uent Finished water Filter effluent Raw water Filter effluent Filter ef3uent

19,275 20,560 23,550 40,000 4,000 28,000 80,800

Gal.

Organic Material Extracted Grams P.p.b.b 1.275 17.5 1.575 21.0 1.795 21.0 9.3 1.420 32.0 0.488 1 0.7 1.160 13.0 4.000

Metered flow of water through Cliffchar carbon. Parts per billion.

solution is then filtered and the residue is washed with four 5-ml. portions of anhydrous ether. The residue is discarded. The ether solutions are combined in a tared flask and gently evaporated in the hood. The flask and contents are dried in a vacuum desiccator over calcium chloride and weighed. The net weight represents the total ether-extractable material found on the carbon. The ether-extractable material is now subjected to group separation, which is effected according to a modification of the method of Shriner and Fuson ( 3 ) . The dry ether-soluble residue is taken up in 25 ml. of ether, placed in a se aratory funnel, and extracted using four 10-ml. portions of 5 6 hydrochloric acid in which is dissolved 20% sodium chloride. The hydrochloric acid-sodium chloride extracts are combined and rendered alkaline with 5% sodium hydroxide solution. The resulting mixture is extracted with four 10-ml. portions of ether. The ether layers are combined and dried over anhydrous sodium sulfate. The dried ether solution is filtered to remove the sodium sulfate. The ether is evaporated and the residue is neighed after desiccation over calcium chloride. The residue is composed of organic basic compounds, largely amines. The basic water solution now contains the amphoteric compounds. It is carefully neutralized with acetic acid and extracted with five 10-ml. portions of ether. The ether extracts are combined and dried as previously. The ether is evaporated and the residue is desiccated over calcium chloride and weighed. This residue is coniposed of amphoteric compounds. The ether layer remaining from the hydrochloric acid eutrac-

The neutral residue so obtained is usually a complex mixture. No general method of analysis is available, but certain methods achieve remarkable separation in many instances. The methods commonly resorted to are fractional crystallization, distillation, and subliniation. When separation is completed, the individual components may be further classified. The sodium hydroxide-sodium chloride solutions are eombined, cor1 d , and saturated with gaseous carbon dioxide. The carbon dioside-saturated solution is now extracted five times with 5-ml. portions of ether, dried, and evaporated. The residue is weighed after drying and is composed largely of phenolic compounds. The sodium bicarbonate-sodium chloride solution is now gently heat,ed and stirred to remove any dissolved ether. It’ is then cooled, made acid to litmus with dilute hydrochloric acid, and extracted with four 10-ml. portions of ether. After drying and evaporation of the ether, the residue is weighed. This residue is composed of st,rong acids and highly acidic phenols. Throughout the ivhole procedure semimicro or micro equipment is used wherever possible. Sodium chloride is added to all aqueous extractants to decrease the solubility of the ether and generally to render more quantitative separations. All volumes are kept as low as possible. Because of the small amounts of materials involved, thorough washing and shaking in extractions are necessary. When separation of the groups is completed, a micro qualita-

ETHER EXTRACT- 2 5 CC.

Add onhydrous No2 S O 4 Allow to stand 48 hours,then filter

t

REStDUE Discard

ETHER SOLUTION

Evaporate ether Weigh,redissalve in ether

t

Extract with 5% H C I - 2 0 X N a C I

++ ETHER SOLUTION

AQUEOUS A d 0 SOLUTION

Extract with 5 % Na OH-20XNa C1

Add No OH solution Extract with ether

A

ETHER SOLUTION Dry with Nap SO Evaparate ethe:

NEUTRA:

~COWOUNOS

ETHER SOLUTION D r y with Na,SO, Evaporate ether

1

PHENOLIC COMPOUNDS Figure 3.

AQUEOUS BASIC SOLUTION Coal-Saturate solution with GOz Extract with e t h e r

I

+

AQUEOUS SO LUTlON

H e a t to remove ether Acidify E x t r a c t with e t h e r Dry’wtth No S O Evaporate‘eth&

ACIDIC C O ~ P O U N D S

f ETHER SOLUTION D r y with No, SO4 E v a p o r a t e ether

BASIC C0)POUNDS ETHER SOLUTION Dry with Nap S O 4 E v a p o r a t e ether

AQ U E0 US ‘SO LUT I0 N Add acetic acid to exact neutrality and e x t r a c t with e t h e r

1+

AQUEOUS SOLUTION Discard

1

AMPHOTERIC COMPOUNDS

Separation of Mixture of Organic Compounds Modified 4ccording to the Method of Shriner and Fuson

1163

V O L U M E 23, NO. 8, A U G U S T 1 9 5 1 tive test for the elements nitrogen, sulfur, and chlorine is made. Nitrogen and sulfur are determined by the method of Feigl ( I ) . RESULTS

Carbon (Cliffchar) filters were placed in service a t six water plants along the Ohio River from Midland, Pa., to Louisville, Ky., during the spring of 1949. An additional carbon filter was placed in service at, Columbus, Ohio (Scioto River). After a metered volume of water had passed through the carbon filters, the filters were returned to the laborat'ory and the carbon was subjected to the treatment' described above. Pertinent dat,a relative to the sampling locations, period of sampling, quantity of water passed through the filters,, and the total quantity of organic matter extracted from the carbons are presented in Table 111. The total amount of organic material extracted, as shown in the last column of Table 111, varied from 9.3 to 32 parts per billion. These values may be subject to significant error due to variable filtration rates and metering errors. The extract from the Columbus, Ohio, sample contained appreciable quantities of organic compounds. This supply is relatively free from industrial pollution. Cincinnati tap water \vas passed through a filter containing Suchar C-190 carbon. The ether extract from the carbon was deep yellow in color. On evaporation a black viscous residue of foul odor 15-as obtained. Elemental analysis of this residue indicated the presence of nitrogen, sulfur, and halogen. The residue was then suhject.ed to qualitative group separation according to t'he schrme shown in Figure 3. The hydrochloric acid-sodium chloride extraction was cont,inued until most of the yellow color mas extracted. The sodium hydroxide-sodium chloride-water extraction yielded three layers. The top ether layer was colored amber, while the lower aqueous layer changed to an orange color indicative of phenolic materials. Between t,hese tvio layers a dark oily middle layer of a tarry vonsistency was noted. The three layers were separated and the tarry mixture wa3 recovered. The middle layer was found to be soluble in very dilute sodium hydroxide but insoluble in 5 % or stronger sodium hydroxide as well as in water, hydrochloric acid solution, and petroleum ether, It \vas soluble in ethyl alcohol, benzene, carbon disulfide, and ether. Crystallization could not be induced from any of these solvents or mixtures thereof. The compound was purified by dissolving in very dilut,e sodium hydroxide and reprecipitating with dilute hydrochloric acid. After several such reprecipitations a light brown amorphous powder \vas ohtained.

Table IV. Lowest Concentration of Organic Extract Detectable by Water-Dilution Method of Odor Testing Source of K a t c r a

Lowest Concentration Detected b y Odor Test P.p.b. 13

Huntington, \V,Ta. Wheeling, Va. 2.5 Loriisvillc, k y . 13 Columbus, Ohio 3 Pomeroy, Ohio 6 Cincinnati. Ohio 2.5 Organic groups Phenolic 30 Xeutral 11 Organic acids 160 a All sources of water were filter effliients before chlorination, except a t Cincinnati, where water treatmenti ncludes brcakpoint chlorination.

"1.

I t \vas thought that this substance might he a complex phenol and that the anomalous behavior was caused by the sodium salt. Subsequent tests, however, did not indicate the compound to be of phenolic nature. The compound could not he acetylat,ed, and on reaction wit,h Gibbs reagent gave a negative response. Using the Lieht.rInarin-Storch reaction a posit,ive violet color was obt'ained, indicating a rosin-type compound. A t'erpenelike odor was noted.

A positive test for chloriiic \viis ol)t,aincd in the pheiiolic group. Reaction with the Gibbs reagent gave a green color indicative of chlorinated phenols. Separation of the individual phenols was not attempted. The phenolic groups as separatcd had t,he typical phenol odor. The acid groups mere charact,erized by an odor similar t'o caproic acid. The group was isolated as a viscous, almost solid mass of bro\vn color. Microdistillation yielded a residue of very high boiling point,, which had the general properties attributed to fatty acids of high molecular weight. It is conjectured that these acids may have originated in the large amount of laundry wastes dumped into the river. The basic group had what is best described as a tobaccolike odor. Pyridine or pyridine-type compounds were shown to be present, using the micro test of Feigl ( 2 ) . The organic neutral compounds appear to be the largest group of compounds isolated. This group is rharxcterizrd hy a particularly offensive odor. KO identifications of the coristituents of this group have as yet been attempted. However, t'he pine splinter-hydrochloric acid test was positive 011 the ether extract. The red color formed is indicative of pyrrole or pyrrole derivatives. Their presence in the neutral groups would in many cases be consistent with solubility considerations. In the analysis of a different filt,er which was iii operation along the Ohio River below a refinery, relatively large amounts of hydrocarbons were isolated from the neutral group. It v-as subjected to niicro steam distillation and the distillate was extracted with ether. Upon evaporation of the ether and solution of t,he residue in fuming sulfuric acid, a floating layer resulted which is highly indicative of aliphatic hydrocarbons. The amphoteric group is characterized by little or no odor. This is consistent wit,h t.he presence of more than one polar group. No separation or identification of this group has been at'tempted. POSSIBLE APPLICATIONS OF TECHNIQUE

Taste and Odor Studies. Tastes and odors in water supplies may arise from industrial and domestic wastes or from natural causes, such as the growth and decay of aquatic plants and animals. In any case, most of the compounds causing objectionable tastes and odors are organic in nature. The concentration, extraction, and separation procedures presented in this paper provide a method for obtaining sufficient quant.ities of the organic compounds for study. By actual taste and odor tests the organic groups or individual compounds most likely to cause tastes and odors may be determined. The effect of these organic compounds may be studied on water-treatment processes, natural purification in streams, and removal from a-ater or wastes. Threshold concrnt rations of taste and odor were determined on the tot,al organic extract, of certain samples and on t.he various organic groups srparatcd (Table IV). Examination of Table IV reveals that 3 to 25 p.p.b. of the total organic extracts from the various watcrs gavc tlrtertable odors when tested in the la,borator?- by thr watrr dilution method. No attempt has been made in this study to compare the threshold odor conrentrations of the various organic extracts with the actual odors occurring in the water supplies from which the extracts were obtained. Carbon filtcrs wcre purposely set up to receivr the rapid sand filter effluents a t t,he water plants, to avoid the possible effect,sof chlorination on the taste- and odor-producing substances. The extract from the Columbus, Ohio, water supply, R hich is not subject, to industrial pollution, produced detectable odor in a concentration of 3 p.p.b. The extract obtained from Cincinnati tap water produced detectable odor a t a concentrat,ion of 25 p.p.b. This water had undergone complete treatment, including breakpoint chlorination. The phenolic, neutral, and organic acid groups which were separated from the Cincinnat,i tap water extract were tcstcd for

1164

ANALYTICAL CHEMISTRY

odor-producing characteristics. The neutral group was detected at a concentration of I1 p.p.b., and the phenolic and acid groups at a concentration of 50 and I60 p.p.b., respectively. Tracing Industrial Wastes. Isolation of the organic compounds gives a clue as to the type of industry responsible for contributing the waste material. By this means a contributor of a given waste may be identified. The method has also proved useful in the direct examination of certain industrial wastes. Study of Reactions in Dilute Solution. The concentration method is useful in the study of certain reactions in dilute solution. The reactions of chlorine and phenol in dilute solution are being studied in this laboratory. T o obtain a sufficient amount of the reaction product for study, large quantities of the dilute solutions are treated with active carbon. The carbon is removed by filtration, followed by ether evtraction and recovery of the product formed in the reaction. These chlorinated phenolic products have been isolated and their identification is being investigated. Other reactions that may be studied include the

products formed by use of chlorine dioxide and ozone when added to various compounds present in water. Possible Basis for Physiological Study. The physiological effect, if any, of small concentrations of organic compounds in drinking water has not been studied. The method presented is helpful in providing sufficient material for studies of a physiological nature. LITERATURE CITED

(1) Feigl, F., “Qualitative Analysis by Spot Tests,” pp. 313-17.

Kew York, Elsevier Publishing Co., 1946. ( 2 ) Ibid., pp. 392-3. ( 3 ) Shriner, R. L., and

Fuson, R. C., “Systematic Identification of

Organic Compounds,” 2nd ed., pp. 252-3, Kew York, John Wiley & Sons, 1940. RECEIVED October 31, 1950. Presented before the Division of Water, S e a age, and Sanitation Chemistry at the 117th Meeting of t h e AMERICAN C H E x l C h L SOCIETY. Detroit, hIich.

Determination of Dipentaerythritol in the Presence of Pentaerythritol JOSEPH H. JAFFE

AND

SHRAGA PINCHAS

Weizmann Institute of Science, Rehovoth, Israel

RIEDRICH and Brun ( 2 ) have shown that the pentaerythFritol prepared by condensation of formaldehyde and acetaldehyde in the presence of calcium hydroxide is always accompanied by a certain and sometimes an appreciable proportion of dipentaerythritol which cannot be removed by any simple method. As for some purposes, only limited quantities of dipentaerythritol are permissible, a quick and easy method for the estimation of the latter in a mixture is important. The methods of Kraft ( 3 ) and Ryler ( 6 ) ,in which the pentaerythritol is determined quantitatively, can (1) be used for the estimation of dipentaerythritol only if one is sure that no other impurities are present in the sample of pentaerythritol, while the chromatographic method suggested by Lew and coworkers ( 4 ) is somewhat lengthy. An attempt was therefore made to estimate dipentaerythritol in the presence of pentaerythritol by infrared spectroscopy. As both substances dissolve in nonpolar solvents with great difficulty and polar solvents are generally unsuitable for this purpose, the absorption of the corresponding acetates-namely, pentaerythrito1 tetraacetate (I) and dipentaerythritol hexaacetate (11)-was measured in carbon tetrachloride solution.

sorption band due to this group has been reported ( 5 )a t about 1120 em.-’ This band was found for I1 a t 1115 cm.+ (Figure 1); a neighboring band of the carbon tetrachloride may somewhat distort it a t very low concentrations of 11. At this frequency, I has a weak general absorption, but there is no trace of a band. Synthetic mixtures of I and I1 were therefore prepared, and it was shown that even in the presence of much I the optical densities a t 1115 cm.-’ permitted a reasonably good estimation of I1 in the mixture. It is suggested that dipentaerythritol be determined in crude pentaerythritol by acetylation (1) and determination of the optical density of a dilute solution of the total resulting acetates in carbon tetrachloride a t 1115 em.-’ The range investigated is 10 to 70% dipentaerythritol in pentaerythritol. For quantities of the contaminant much smaller than IO%, it will probably be advisable to enrich the dipentaerythritol in the sample before applying the method of assay described. Any higher “polymer” of pentaerythritol will be determined by this method as dipentaerythritol. EXPERIMENTAL

I1

CH3.COOCH2

CHpOOC.CH3

CH3.COOCH2

CH?OOC.CHB

\ / CH,.COOCHz-C-CH2-O-CH2-C-CH,OOC.CH, / \

The only significant differences between the spectra of these compounds were to be expected in the regions of the absorption of the HpC-O-CHz linkage, whirh nrrurs in I1 only. -4n ah-

The infrared absorption measurements were made xith a Perkin-Elmer, Model 12 C spectrometer, using a rock salt prism and a slit width of about 0.4 mm. In order to minimize interaction between I and 11, very dilute solutions (total concentration of about, 0.01 gram per ml.) were used and a cell of 2-mm. thickness was constructed so as t o obtain a sufficient absorption. The cell consisted of a pair of rock salt plates, between which a lead ring was amalgamated as spacer. Samples were introduced by means of a hypodermic syringe needle, through two small holes drilled in the spacer. During the measurements, these holes were closed with pins in order to reduce evaporation of the solvent. In the quantitative work, it was found more accurate t o set the spectrometer a t 1115 cm.-I, the group, than to measabsorption maximum of the H2C-O-CH2 ure the complete absorption spectrum. A blank measurement with the pure solvent was carried out immediately before that with the solution, and the difference was taken as the optical density of the solution.