Identification of an organic compound leached from a membrane filter

584

SHORT

Talanta, Vol

24. pp. 584-586.

Pergamon

COhUvfUNlCATlONS

Press. 1977. Prmted 1x1Great Brltain

IDENTIFICATION

OF AN ORGANIC COMPOUND FROM A MEMBRANE FILTER

LEACHED

AKIRA OTSUKI and KEIICHIRO FUWA Division of Chemistry and Physics, National Institute for Environmental P.O. Yatabe, Tsukuba, Ibaraki 300-21, Japan

Studies,

(Received 11 January 1977. Accepted 11 February 1977)

Summary-An organic compound leached from a Millipore HA filter has been identified as the nonionic surfactant polyoxyethylene nonylphenyl ether with a degree of polymerization of 4-10 for the ethylene oxide unit. It is suggested that the Millipore HA filter should be used only after several rinses with sample or doubly-distilled water.

Membrane filters are widely used in the fields of microbiology, limnology, oceanography and environmental analytical chemistry in order to remove suspended matter from water samples. Although sample contamination resulting from membrane filters has been recognized for several years, it has not been known what kind of compounds may be leached from the membrane filter during filtration. Even if the amount of the compounds leached from the filter is extremely low, it should be known for further experiments. In an attempt to develop a direct method for determination of phthalate esters in water, one of the authors has examined the reverse-phase adsorption chromatography of phthalate esters in aqueous solution, using a high-pressure liquid chromatograph with solvent programmer.’ The order of gradient elution of the phthalate esters adsorbed from water on an octadecyltrichlorosilane-bonded bead column was found to be directly related to the alkyl carbon number in the ester group. In the course of studies on the application of this method to environmental water samples, an interfering compound having similar chromatographic behaviour to the phthalate esters was found in lake water samples filtered through a Millipore HA membrane, and experiments done with a glass-fibre filter combusted at 450” for 4 hr indicated that the interfering compound was leached from the Millipore HA filter. We have tried to identify the compound because it was often present in sufficient amount for its absorption maximum at 224 nm to be directly observable in the spectrum of the filtrate when 10 ml of doubly-distilled water were passed through a Millipore HA filter 47 mm in diameter. EXPERIMENTAL

Apparatus

A model ALC 202 liquid chromatograph equipped with two 6000A pumps, a model 660 solvent programmer, and a model 440 UV detector operating at 280 nm (Waters Associates, USA) was used. A 60 cm x 9 mm column. packed with Bondapack C-18 (Waters Associates, USA), was used. Elution was started with 100°/Owater and continued to 100% methanol with a hyperbolic gradient. “Curve 2” of the Waters programmer was selected. The setting time for the programme was 15 min. Field-desorption mass-spectra were obtained on a JEOL JMS-OlSG-2 double-focusing high-resolution mass-spectrometer with a combined field-desorption/field-ionization/ electron-impact ion-source (Japan Electronic and Optics Laboratory). Tungsten wire (10pm in diameter) on supporting spots was conditioned in a similar manner to that

described by Schulten and Beckey,’ with a model MSFDA01 activator (Japan Electronic and Optics Laboratory), and used as an anode. The anode potential was 8.0 kV and the cathode potential - 5.0 kV. The emitter current was 12 mA. The sample was applied to the conditioned emitter by the dipping technique.’ A Hitachi 323 spectrophotometer and a Hitachi 285 IRspectrophotometer were used. Doubly-distilled water was passed through the Millipore HA filter (47 mm in diameter) and pumped through the column at 3 ml/min for 4 hr to retain the compound of interest on the column, then gradient elution was started after the flow-rate had been changed to 2 ml/min. The fraction eluted at a mobile-phase composition of about 5% water/95% methanol was collected. This was repeated and the combined fractions were concentrated under reduced pressure at room temperature. Two hundred filters from 5 boxes were used for this study. The methanol used was “nanograde” reagent (Mallinckrodt, USA). RESULTS

AND DISCUSSION

The ultraviolet spectrum of the fraction is shown in Fig. 1. Absorption peaks at 224 and 276 nm indicate the pressure of a benzene ring. Figure 2 shows the field-desorption mass-spectrum of the concentrate, suggesting that the lowest molecular weight of the compound is 396 and that it partly has a polymer structure based on a monomer with m/e = 44. The elemental composition of this monomeric unit should be COZ, C2H40, N20, CH4N2 or CzHF because of the “nitrogen rule”. The possible structures are as follows :

-TV-%-,

-KHr--CHr-O),-,

--(T-CH2).-_

0

OH (I)

(II)

(III)

-+H-O),->

--(N)-,

--(I;-NH).,

CH3 (IV)

NO

CHJ (VI)

-(CH-NH).-, NH2 (VII)

(V)

--(CH=CF).-

(VIII)

585

SHORT COMMUNICATIONS

4 5 /

I

rc)

I

3000

8

2000

I

,

1500

1000

1

50(

cm-’

f

cl

Fig. 3. Infrared spectrum of a thin layer on a KBr plate.

On the basis of the information two structures were suggested:

above, the following

C&Hz,+ I . CbH4. O(CH~CH~O)~H (IX) Wavelength,

nm

or

Fig. 1. Ultraviolet spectrum of the fraction containing the organic compound.

C#z,+

I .C6H4. O(CHO),H I CH3

(w However, (V) and (VI) are the most unlikely chemical structures of all those suggested. If the compound partly has structure (VIII), the absorption maximum at 275 nm would be shifted to a longer wavelength of more than 290nm with increasing number of ethylenic groups4rs Because it is soluble in water and has hydrophobic benzene ring and alkyl chain groups, it seems unlikely that the monomeric unit in the remainder is the hydrophobic C2HF of structure (VIII). Figure 3 shows the infrared spectrum of a thin layer on a KBr plate. The strongest peak with two shoulders at about 1100 cm-” suggests the presence of an ether bond, supporting the structures (I), (II) or (IV). The weak broad peak at 3100-3500 cm-’ excludes the structures (III), (VI), and (VII). The peaks at 1500 and 1610 cm-’ confirm the presence of the benzene ring and there is no evidence as to 1,2 or 1,3 disubstitution. There appears to be a substantial amount of methyl and methylene groups as evidenced by the relatively strong peaks near 2900 cm-‘. This agrees with the behaviour in the reverse-phase adsorption chromato~aphy in the isolation process.

4

,

440

528

Li 396

1 400

500

m/e

Fig. 2. Field-desorption

600

700

mass-spectrum of the concentrate.

The structure (IX) may correspond to polyoxyethylene alkylphenyl ether, which is commercially available as a non-ionic surfactant. Since there is no peak less than m/e = 396 in Fig. 2, the values of m and n that satisfy a rnol~ul~ weight of 396 can be calculated. Only when m is 9 and tl is 4, this molecular weight is obtained to give the elemental composition CZ3H4c0s. if there has been no fragmentation of the compound in production of the mass spectrum and the elemental composition is correct, the mass spectrum can be completely interpreted in terms of increase in the molecular weight by 44 with unit increase in n. The infrared spectra of the compound obtained, and of Igepal CO-880 and Antarox CO/850 (both of which are polyoxyethylene nonylphenyl ether),6 were compared and found to be essentially identical, although the degree of polymerization of ethylene oxide in the latter two is different7 Polyoxyethylene octyl-, nonyl- or dodecylphenyl ether is widely used as a wetting agent, penetrant and detergent. It may therefore be concluded from these results that the compouitd was ~lyoxyethylene nonylphenyl ether with a degree of polymerization of 4-10 for the ethylene oxide unit. The amount leached from a 47-mm diameter HA filter was estimated from the benzene-ring absorbance at 224 nm by use of Igepal CO-880 as standard. When lo-ml of doubly-distilled water was passed through a Millipore HA filter 47 mm in diameter, the filtrates showed three types of absorption spectrum: one clearly had the peak at 224 nm, the second just a shoulder, and the last no peak at all. The results indicated that 100-200 1.18of the polyoxyethylene nonylphenyl ether would have been leached from one 47-mm filter when the filtrate spectrum had a peak at 224 mn. This amount would be extremely im~rtant in the me~urement of low levels of dissolved organic carbon in natural waters, such as sea-water, depending on the sample volume to be filtered, and when micro-organisms in the filtrate are to be cultured.

586

SHORT

COMMUNICATMlNS

The present study suggests that the membrane filter should only be used after several lo-ml rinses with doublydistilled water or sample. Acknowledgement-The authors thank Mr. K. Nojima of Japan Electronic and Optics Laboratory for his assistance in the ~eld~~o~tion mass-spectrometry.

2,500. 4. K. W. Hausser, R. Kuhn and A. Smakula, 2. Phys. Chem, 1935, B29, 363. 5. S. Akiyama, Y. Takeuchi, A. Yasuhara, M. Nakagawa and K. Nishimoto, Bail. Chem. Sot. Japan, 1973, 46, 2830. 6. D. Hummel, ~den~~~ca~~onand Anatysis of Surface Active Agents by infrared and Chemical Me~hods~ Spectra Fume, Spectrum 344, Interscience, New York,

REFERENCES

A. Otsuki, J. Ckromatog., 1977, 133, 402. 2. H. R. Schulten and H. D. Beckey, Org. Mass Spectry., 1972, 6, 885. 1.

Tolanio. Voi

3. H. D. Beckey, Intern. J. Mass Spectry. Ion Phys., 1969,

1964. 7. M. J. Rosen and II. A. Goldsmith, Systematic Analysis of Surface-Active Agents, p. 491. Wiley-Interscience, New York, 1972.

24, PP 586-S-588 Pergamon Press, 1977 Prmted m Great Bntam

STUDIES ON THE POLAR~RAPHIC REDUCTION OF THE AZOMETHINE BOND IN THE PRODUCTS OF ~-KETO-ESTERS COUPLED WITH ARYL~IAZONIUM CHLORIDES WAHID U. MAX.IK*, R. N. GOYAL@ and RAJEEV JAIN Department of Chemistry, University of Roorkee, Roorkee, India (Received 23 August 1976, Revised 18 November 1976. Accepted 9 March 1977)

Summary-The products of coupling B-keto-esters with aryldiazonium chlorides have been studied polarographically and give a single well-defined 4-electron diffusion-controlled irreversible wave in the oH range 2.0-11.0. The effect of electron-donatine and electron-withdrawinr substituents and the correlation -between the half-wave potential and Hammett substituent cons& have been studied.

The polarographic behaviour of compounds having an -N=Nlinkage has long been the subject of investigation but that of compounds likely to have a --C==Nlinkage in their tautomeric form has attracted little attention.ls3 The products of coupling b-keto-esters with diazotized aromatic amino-compounds exist in tautomeric forms, which possess these two types of group: COR /

Ph-N=N-CH

\

COR Ph-NH-N&

\

COOR

(A)

/ COOR

0%

Since these products have been extensively used as intermediates in the preparation of a large number of biologically important heterocyclic compounds, it was considered worthwhile to seek a polarographic criterion for distinguishing between the tautomers. EXPERIMENTAL

Reagents The products listed in Table 1 were prepared according to the literature procedure.4 Stock solutions (1 x lo-” M) of the products were prepared in methanol. Apparalus

A Cambridge pen-recording polarograph was used. The capillary had a drop-time of 3.24sec and mercury flow * Present address: Vice Chancellor, Bundelkhand versity, Jhansi.

Uni-

of 2.68 mgisec at h 30 cm, in 1 M potassium chloride. All measurements were made at 15 + 0.1”. Procedures

Britton-Robinson buffer solution (7.5 ml), 1 M potas sium chloride (1 ml) and freshly prepared 0.1% gelatin solution (0.5 ml) were mixed and dearated by passage of purified nitrogen for about 2 min. The solution was then mixed with 1 ml of the stock solution of a coupled product and deaerated for a further 3min. Polarographic curves were then recorded. Controlled-potential electrolysis was carried out by the method of Devries and Kroon’ in a small volume (0.3 ml) in a modified H-cell, with a mercury-pool electrode. The temperature coefficient was calculated by the method suggested by Nejedly.6

RESULT

AND

DISCUSSION

For compounds I-VII and XI-XIX single four-electron waves were observed over the pH range 2-11. For compounds VIII-X this wave was followed by another fourelectron wave at a much less negative potential, over the entire pH range. The heights of the waves were approximately the same for all the substances (Table 1). All the wave-heights were found to be pH-independent in the pH range 2.0-11.0 and the diffusion-controlled nature of the limiting current was established in the usual way. The low values of the tem~rature coefficient (1.~1.6~~deg) also indicated the effusion-controlled nature of the waves. The half-wave potential of all these substances was found to become more negative with increase in pH in