Table I. mination
Sample
Sulfur ancl Chlorine Deterfrom Neutron Activation Analy!;is"
is, p.p.m. Found Found C1, Added via #Sa5 via Psn p.p.ni.
20.2 0.1 0.1 carried out by the Activation Analyigis Service of the General Atomic Division of General Dynamics Corp., P. 0. Box 608, San Diego 12, Calif. 24.6 55.6 C 52.4 These analyses were
A B
0.86 0.09 0.009
2.7 0.27 0,029
Chlorine was determined by irradiation of 5-ml. aliquots of sample solutions for 500 kw. hours at :t thermal-neutron flux of 1.8 X 10l2 neutrons cm.-z second-' in a TRIGA Mark I reactor. Chlorine was separated as AgC1 and detected instrumentally using the 1.65m.e.v. ?-ray of 37-minute Cla8. The slow neutron capture reaction involved in thi- case was: Clm(n,,)C138 RESULTS AND DISCUSSION
The specific activities of the sulfur and phosphorus fractions were 1.0 c.p.m. per Ng. of S and 7 5 c.p.m. per pg. of P, respectively. Results of the study are summarized in Table I. Discrepancies exist between the amount of sulfur added and the amount found. Determination of sulfur by S 3 5 was both inconsistent and inaccurate (due to radioactive contaminants in the BaS04 precipitates), whereas
when sulfur was determined by P32 there was a discrepancy but it was a consistent one. High results for sulfur may have been caused by significant but variable removal of sulfonate from the ion-exchange resin during deionization of the water and, because of this, a weak carboxylic resin should be considered for this application. More likely, a small amount of phosphorus impurity in the sulfonate might have caused the discrepant values found by P3?. If the sulfonate contained about 0.5yoP, the results would be in exact agreement. The Pa* activity varied linearly with the amount of sulfur added, but the amounts of sulfur indicated by the Pa2values were about three times the amounts of sulfur originally added. Because of its low concentration, chlorine would not have caused any significant interference (0.1 p.p.m. C1 would only be equivalent to 0.007 p.p.m. S, determined as Pa2). -4slittle as 0.01 p.p.m. of sulfur can be detected in water, provided chlorine and phosphorus have been removed to eliminate interferences from the Pal ( 7 ~ , y ) Pand ~ ~ C P ( n , ~ u ) Preactions. ~~ The most sensitive reaction for determining sulfur is S32(n,p)P32. In a previous study, it was indicated that for sulfur to phosphorus ratios greater than 10, both elements may be determined by counting Pa2with the use of a double irradiation technique a t different fast-to-slow neutron fluxes (1). Similar techniques may also be fruitful for total S and total P in water. The method also shows promise for determination of alkyl-benzenesulfonate S
when this is the only source of S; limitations of the methylene-blue method for this determination have already been reviewed (8). To determine both total sulfur and each specific amount of ionic species in water, activation analysis must be combined with a recent method for differentiation among sulfate, sulfite, and sulfide by anionic chromatography ( 3 ) . ACKNOWLEDGMENT
I am indebted to Vincent P. Guinn, General Dynamics Corp., for his helpful suggestions in preparation of this paper. LITERATURE CITED
( 1 ) Bouten, P., Hoste, J., Anal. Chim. Acta 27,315-9 (1962). (2) Gibbons, D., Simpson, H., R1CC/7, Conf. on Use of Radioisotopes, Copenhagen, September 1960. (3) Iguchi, A., Bull. Chem. SOC.Japan 31,600-5(1958). ( 4 ) Lauman, C. W., N. Y. State Water
Pollution Control Board Res. Rept.
90.6 , June 1960.
(5) Leddicotte, G . W., Emery, F., U. S. htomic Energy Comm. Rept. ORNL-2453,23 (1957). ( 6 ) Nichols, M. S., Koepp, E., J . Am. Water W o r k s Assoc. 303-6 (1961 ). (7)Walton, G., Zbid., 52, 1354-62 (1960). (8) Wayman, C. H., U. S . Geol. Survey Prof.Paper 450-B,B117-B120 (1962).
COOPER H. WAYMAN C. S. Geological Survey Denver, Colo. RECEIVED for review Sovember 19, 1962. Accepted February 27, 1963. Publication authorized by the Director, U. S. Geological Survey.
Rapid Paper Chromatographic Analysis of Phosphate Mixtures SIR: -4 new singk-step ascending paper chromatographic method was developed for the rrtpid analysis of commercial phosphates and phosphate mixtures for the fcdlowing species : ortho-, pyro-, tripoly,, trimeta-, and higher polymeric phosphates. Since the introduct .on of ascending paper chromatography into the field of condensed phosphzte analysis by Ebel and Volmar (3) a major effort has been made t o improve this analytical tool in simplicity, accuracy, and speed. A summary of this wo:k was published by Hettler ( 4 ) . A noteworthy effort to improve the speed of the method was made by Bernhart and Chess (1) who used a densitometer to measure directly the concentration of the separated phosphates on the paper chromatogram.
This method suffers in accuracy because of the difficulty of completely hydrolyzing the various polyphosphates without destroying the paper (6). An improvement in the hydrolysis technique was reported by Kolloff (6), but evaluations in this laboratory indicate that the severe hydrolysis conditions used often cause embrittlement of the papers, making subsequent handling difficult. In the method developed by KarlKroupa ( 5 ) ,the hydrolysis is completed after the separated phosphate species are eluted from the paper and analysis is made by the colorimetric method of Martin and Doty (7'). However, this method has the serious disadvantage that samples containing both tripolyand trimetaphosphate require two separate chromatographs, one of which is
two-directional with two separate solvents. By combining the best features of these methods, with modification, a method has been developed which is both rapid and accurate. EXPERIMENTAL
Reagents and Apparatus. The reagents and apparatus are the same as those described by Karl-Kroupa ( 5 ) with the following exceptions: Chromatographic solvent. Dissolve 25 grams of trichloroacetic acid in (concd.) water, add 1.75 ml. of ",OH and dilute with water to 175 ml. Add this solution to 325 ml. of acetone and mix. This solvent can be used for 4 papers only, and should be freshly prepared each day. VOL. 35, NO. 6, MAY 1963
769
ACKNOWLEDGMENT
Table 1.
Precision Data
95% confidence level Phosphate species Ortho
Pvro Sipoly Trimeta High poly
.4v. of all
determinations 0.29 5.07
92.18 2.08
0.36
Refrigerator, or cold box, niaintained between 10" and 15" C. The chromatographic paper is cut into 9- X 8-inch sheets, and 7 rather than 8 sample spots are marked for each sample. Procedure. The procedure is the same as t h a t described by KarlKroupa except that seven 5-,ul. droplets of sample solution are used, the jar containing the chromatographic solvent is placed in a refrigerator a t 10' to 15" C. for a t least 15 minutes before introducing the paper, and the solvent is allowed to ascend 6','* to i inches
Standard deviation 3Z0.07
3Z0.25
& O . 46 10.16
*o.
10
Single 0.15 0.57
0.98 0.34 0.21
Av. of
duplicates 0.11 0.38 0.70 0.24 0.15
from the bottom while in the refrigerator. The time required depends on the temperature and the freshness of the solvent and should be between 60 and 120 minutes. Aifterdevelopment of the chromatogram, the trimetaphosphate hand is clearly separated from the tripolyphosphate and is marked and cut out separately, thus giving five separated fractions instead of the four obtained by the Karl-Kroupa method. The precision data in Table I were obtained from 22 analyses on eight ;ample5 of sodium triphosphate.
The assistance of L. T. Troost in obtaining much of the analytical data is appreciated. LITERATURE CITED
(1) Bernhart, D. Pi., Chess, W.B., AXAL.
CHEX.31, 1026 (1959). ( 2 ) Bernhart, D. N., Chess, FV. B., Victor Chemical Works, Chicago, - 111..
private communication, 1960. (3) Ebel, J. P Volmar, Y., Compt. Rend. 233, 415 (1951). ( 4 ) Hettler, Hans, J . Chromatog. 1, 389 (195s). ( 5 ) Karl-Kroupa, Editha, ANAL. CHEW
28, 1093 (1956). (6) Kolloff, -4. H., Ibid., 33, 373 (1961). (i) Martin. J. B.. Dotv. " , D. AI.. Ibid.. 21,
965 ('1949). '
K. J. FECHS F. !V. CZECH Inorganic liesearch 'and Development Department F l I C Corp. Carteret, S . J.
Nonaqueous Titration of p-Nitrophenylhydrazones with Tetra butylarnrnonium Hydroxide SIR: Prior to this study it 11-as generally considered that the condition required for the nonaqueous titration of nitroaromatic amines as acids was the presence of at least two nitro groups or one nitro group and one or more halogen atoms in the 2:4 or 2:4:6 positions. On this basis Fritz, Moye, and Richard (3) reported the titration of a number of aromatic amines as acids in pyridine with triethyl-n-butylammonium hydroxide. Sensabaugh, Cundiff, and Markunas (6) subsequently demonstrated that the 2,4dinitrophenylhydrazones could be quantitatively titrated as weak acids in pyridine with tetrabutylammonium hydroxide, and Robinson, et al. (4, 6 ) showed that the 3,5-dinitrobenzoates were sufficiently acidic to be titrated as acids with tetrabutylammonium hydroxide. The present work shows that p-nitrophenylhydrazones of aldehydes, ketones, and carbohydrates are also acidic enough to be titrated in pyridine with tetrabutylammonium hydroxide. EXPERIMENTAL
Procedure. Reagents and apparatus have been described in previous publications (1, 2, 5 ) . p - Nitrophenylhydrazone samples were prepared in our laboratories by standard procedures and recrystallized to constant melting points. 770
ANALYTICAL CHEMISTRY
=Iccurately weigh 0.06 to 0 . 0 i meq. of p-nitrophenylhydrazone sample, place in a 250-ml. electrolytic beaker, dissolve in 50 ml. of pyridine, heating if necessary, and titrate potentiometrically under a nitrogen atmosphere with 0.025 tetrabutylammonium hydroxide. Correct the potentiometric equivalence point for solvent blank and calculate the neutralization equivalent using the corrected total volume. RESULTS A N D DISCUSSION
All p-nitrophenylhydrazones titrated potentiometrically as very weak acids. Indicators could not be used because of a deep color which developed immediately upon addition of titrant. Figures 1 and 2 show typical potentiometric curves of the p-nitrophenylhydrazones. Curves in the figures are shifted for clarity. shorter millivoltage range was observed in titration of the carbohydrate derivatives than that obtained with the aldehyde and ketone derivatives. This can be seen from Figure I , curve B , which represents titration of the glucose derivative. -111 derivatives of dicarbonyl compounds titrated as dibasic acids with a very weak first inflection and a final strong inflection as seen from Figure 1, curve A , which represents titration of the 2,3-pentanedione derivative. Derivatives n-hich contain a second titratable acidic group such as the
phenolic hydro.\-1 group of salicylaldehyde or p-hydroxybenzaldehyde might elhibit one or t x o inflections in the potentiometric curve depending upon the position of the acidic group. This is seen from Figure 2, curves -4 and B, which represent titration of the dicylaldehyde and p-hydroxybenzaldehyde derivatives respectively. Table I lists results from titration of representative p-nitrophenylhydrazones of aldehydes, ketones, and carbohydrates. T'alucs listed are the mean of a minimum of two determinations, and \\-ere obtained using a PrccisionDoir- Recordomatic Titrator. Standard deviation for the procedurr using the Recordoniatic Titrator was r0.42. This value was reduced to 1 0 . 3 2 using a manual titrator or equivalent pH meter. -1s n-ith any analytical procedure of this nature, purification of the samples with regard to anomalous side reaction products or trapped reagent is essential. In a fern instances quantitative results were realized only after repeated recrystallizations even though melting points and crystalline structure agreed with the literature. The pyridine solution of pure derivatives changes color immediately upon addition of titrant, usually from yellow to red or green depending on the particular derirative. delay in the developmrnt of color until after scvrral