ty for F- ( a l ) and that for SiFs2- (az) and using the expressions
c - = 6 2 - 4 and c a,f - a ,
SIF,?-
[azf -- ] (1)
- 1 6z
= L!
a2
a2
6i
where f = SiF+ that would result from 1ppm of F-
ACKNOWLEDGMENT Thanks are due to S. Sreepathy for help in regard to literature survey. The authors express their thanks to S. Ramaseshan for his keen interest in this work.
a1
.
Received for review March 28, 1973. Accepted October 29, 1973.
Losses of Trace Concentrations of Cadmium from Aqueous Solution during Storage in Glass Containers W.
G. King, J. M. Rodriguez, and C. M. Wail
Department of Chemistry, University
of Idaho, Moscow, Idaho 83843
To whom correspondence should be addressed.
The role of heavy metals in the environment and their biological impact in the ecosystem have caused considerable attention in recent years. Many heavy metal pollutants, among them particularly mercury and cadmium, have been shown to be highly toxic even in the ppb range. Accurate detection of such metal pollutants in trace quantities has become increasingly important for environmental monitoring programs. One difficulty commonly encountered in the analysis of liquid samples is the loss of heavy metals during storage. Severe losses of mercury (over 90% in ppb range) from water samples have been reported to take place over a short period of time when stored in various containers ( I , 2). Several reports of radioactive isotope studies have also shown that losses of many metal ions from aqueous solution occur by adsorption to container walls (3, 4 ) . In a report on the analysis of traces of heavy metals including cadmium, McFarren and Lishka mentioned the necessity of acidification of solution with nitric acid to pH below 1.0 in order to prevent precipitation and adsorption of the metals on the walls of borosilicate glass bottles ( 5 ) . Posselt and Weber have observed that at pH > 7 the loss of cadmium to borosilicate glass is more severe than other types of containers and that Teflon, polyethylene, and polypropylene have the lowest adsorptive properties relative to aqueous species of cadmium (6). Previous studies have indicated that adsorption of cadmium ions to container walls can take place at high pH, but quantitative data are lacking. The present study was undertaken to determine the magnitude as well as the mecihanism for the loss of cadmium from water samples stored in various containers. Losses of cadmium in solution's of different concentration and pH were measured with respect to time. Radioactive losCd (half-life 470 days) was employed as a tracer for this study.
EXPERIMENTAL The bottles chosen were of the general storage type commonly found in most laboratories. The soft glass bottles were 32-02 (1) (2) (3) (4)
(5)
R. V. Coyne and J . A . Collins,Anal. (:hem., 44, 1093 (1972). R . M . Rosain and C. M . Wai, Anal. Chim. Acta, 65, 279 (1973) 0. E. Robertson, Anal. Chim. A c t a . , 42, 533 (1968). P. Benes and I . Rajman, Collect. Czech. Chern. Cornrnun., 34, 1375 (1969). E. F. McFarren and R. J . Lishka, "Evaluation of Laboratory Methods for the Analysis of lnorganics in Water," Chapter 15 in: "Trace Inor-
ganics in Water." American Chemical Society, Washington, D.C., 1968. (6) H. S.
Posselt and W . J. Weber, Jr., "Environmental Chemistry of Cadmium in Aqueous Systems," Tech Rept. T-71-1, University of Michigan, A n n Arbor, Mich., 1971.
Armstrong's Boston rounds and the borosilicate glass bottles were 500-ml glass stoppered reagent type. Initially, several types of plastic bottles were tried: polyvinyl chloride, polypropylene, linear polyethylene, all by Nalgene, and Plax's polyethylene. Since all of the polymers appeared to give the same results, polyethylene was selected as the representative. Adsorption experiments were started with new bottles which were cleaned by the following procedure: the bottles were given a detergent wash, a distilled water rinse, followed by a dilute nitric acid wash, and finally rinsed several times with demineralized water. The demineralized water was prepared by passing distilled water through a series of ion exchange columns which resulted in the final water having a resistance greater than 18 megohms. The water used to make the 25 ppb and carrier-free samples was the same as that used for the final rinse. No cadmium background was detectable by atomic adsorption for the demineralized water. The pH of the solutions was adjusted using reagent grade nitric acid or sodium hydroxide. The experiments were started by placing 500 ml of sample solution in the bottles to which was added 25 pl of logCd tracer. The bottles were then shaken and a 1-ml control sample was extracted from each. In the experiments with 25 ppb concentration, 1 ml of carrier solution was added to each bottle before the tracer. The carrier solution was composed of 3.44 mg/100 ml of reagent grade Cd(N03)Z - 4 HzO. The monitoring of lo9Cd activity in each experiment was done by taking 1-ml aliquots a t appropriate intervals. Experiments were conducted in duplicate. The overall experimental error was 3% or less. The samples were prepared for liquid scintillation counting by placing the 1-ml aliquots into Kimble disposable glass scintillation vials containing 20 p1 of a saturated cadmium nitrate solution. The purpose of the cadmium nitrate solution (approximately 4 X 10-3N) was to reduce the probability of adsorption of the lo9Cd tracer onto the walls of the counting vials (7). Then 19 ml of scintillation solution was added and the vial shaken to produce a homogeneous mixture. The scintillator solution was of Butler's foqnulation (8) containing 4 grams PPO, 0.1 gram POPOP, and 120 grams of naphthalene in one liter of dioxane. The above chemicals were scintillation grade. The lo9Cd was counted on a Packard liquid scintillation counter with a lower discriminator setting of zero, a 1.1%gain, and an 8% window. The samples in counting vials were allowed to stabilize by remaining in the counter for 12 hours before counting. At least a total of 10,OOO count was collected for the samples. A blank was used to determine the background and a standard sample was used to monitor counting efficiency.
RESULTS AND DISCUSSION Losses of cadmium from aqueous solution during storage appeared to depend upon at least three parameters; container material, pH of solution, and concentration of cadmium. When distilled water samples spiked with carrier-free Chase and J . L. Rabinowitz, "Principles of Radioisotope Methodology," Burgess Publishing Co.. Minneapolis,M i n n . , 1967, p 140. (8) F. E. Butler, Anal. Chern., 33, 409 (1961). (7) G . D.
A N A L Y T I C A L C H E M I S T R Y , VOL. 46, NO. 6, M A Y 1974
771
20 Time (hours)
50
IO
40
50
l i m a (hours)
Figure 1. Per cent of lo9Cd in water during tion of p H of solution. Distilled water with
storage as a funccarrier-free lo9Cd
stored in soft glass containers. 0, pH 10.0; o , pH 9.0; 0 , pH 8.0; A , p H 6.9
lo9Cd were stored in bottles made of polymers, virtually no loss of cadmium activity was observed in the pH range 3 to 10. Of the three types of plastic bottles tested (polyethylene, polypropylene, and polyvinyl chloride), all of them showed less than 3% of loss of lo9Cd activity from the solution even for a period of two weeks of storage. However, when the same solution a t pH 9 was stored in a soft glass container, loss of cadmium activity was found to be as high as 75% after one day of storage (Figure 1).To check the total activity in the system, a soft glass container was broken and various segments were counted. The lo9Cd activity was detected on the sides as well as on the bottom part of the container. The total lo9Cd activity found on the glass walls was comparable to the logCd activity lost from the solution. Apparently, loss of cadmium from distilled water stored in glass containers occurred by adsorption to container walls. The fact that adsorption of cadmium took place only in glass containers seemed to suggest a strong surface interaction between cadmium ions and silicates. Polymer surfaces did not seem to interact with cadmium ions under this experimental condition. Changes in pH had a marked effect on the extent of loss of cadmium from aqueous solution. Figure 1 shows the percentage of cadmium loss at different pH of distilled water samples spiked with carrier-free lo9Cd in soft glass containers. No detectable loss of cadmium activity from the solution was observed at pH less than 7 . The two curves for pH 9.0 and 10.0 nearly coincide and approach a maximum adsorption value of 75% in 16 hours of storage. Furthermore, a linear relationship was observed between log (C - C,) and time for the loss of cadmium at pH 9.0 or 10.0, where C, is the equilibrium concentration of cadmium in solution when t becomes large and C is the concentration of cadmium at any time. This relationship seems to support a first-order kinetics for the loss of cadmium from the solution under these experimental conditions. The rate constant k for this process is in the order of 0.165 hr-I which corresponds to a half-life of about 4.2 hours. The shape of the curve for pH 8.0 (Figure 1) is quite different in that it shows a maximum cadmium loss of about 18% after 8 hours of storage. Perhaps a slow deadsorption equilibrium process was involved in the system a t this pH range. Experiments with borosilicate glass containers shows characteristics similar to those of the soft glass experiments except that the percentage of cadmium loss at each pH in borosilicate glass was found to be about 0.7 times that of soft glass. Our experimental results have confirmed that loss of cadmium to glass containers occurs at pH > 7 and have shown that severe loss of cadmium by adsorption to glass surfaces takes place at pH > 8. During this experiment, it was also noticed that the adsorption of 109Cd activity by the walls of glass container was a revers772
ANALYTICAL CHEMISTRY, VOL. 46, NO. 6 , M A Y 1974
Figure 2. Per cent of lo9Cd in water during storage with respect to the initial concentration of cadmium in solution. Distilled water at pH 10.0 stored in soft glass container. 0, carrier-free lo9Cd; 0 , 2 5ppb; 0 , 100 ppb; A , 200 ppb Cd 1oorn
n
*
Y
Y
90-1' I 0
1
10
20 10 Iine(hourr)
40
50
Figure 3. Per cent of lo9Cd in water during storage as a function of pH of a solution with 25 ppb Cd in soft glass containers. 0, pH 10.0; 0 ,p H 8.6; 0 , pH 8.0; A , pH 6.9.
ible process. When a lo9Cd spiked solution at pH 9 stored in glass container for an extended period of time was acidified to pH 6, the lo9Cd activity in the solution returned to its initial value after one day. It has been shown that in aqueous cadmium hydroxide solution, the Cd2+ species exist as the predominantly stable form of cadmium at pH < 8. The Cd(OH)+ species begin to appear at pH > 7 with its maximum concentration in solution at approximately pH 10. The neutral Cd(OH)ziaq) starts to appear at pH > 8 and reaches a maximum at pH 11-12 (6). In these experiments, the observed adsorption for the pH range of 7 to 10 corresponded to the existence of the Cd(OH)+ and Cd(OH)ziaq)species in the aqueous solution. The increase in cadmium loss from pH 7 t o 10 as shown in Figure 1 is consistent with the increase in Cd(OH)+ concentration. Because of experimental difficulties involved in counting lo9Cd activity in highly basic solutions, we were not able to perform experiments at pH greater than 10. Therefore, the role of Cd(OH)?(,,, in the adsorption process could not be fully determined. Nevertheless, our data appear to suggest that Cd(OH)+ instead of Cd2+ is a reacting species in the adsorption process. The initial concentration of cadmium present in water also affected the degree of loss of cadmium during storage. Figure 2 shows the percentage loss of cadmium as a function of time for distilled water samples at pH 10 with different initial cadmium concentrations. The curve for 25 ppb cadmium reaches a maximum value of about 35% cadmium loss after 20 hours of storage in soft glass container. This is lowered by a factor of 2 compared with the experiments for carrier-free logCd in solution at the same pH. The curves for 100 ppb and 200 ppb cadmium are similar to the 25 ppb curve except that the degree of cadmium loss is less pronounced for the higher concentrations. The dependence of pH on the loss of cadmium from distilled water samples with 25 ppb initial cadmium concentration is shown in Figure 3 . For many environmental monitoring programs, the significant level of cadmium is in the 1-102 ppb range.
There are some obvious differences in the nature of the loss of cadmium relative to the loss of mercury from aqueous solution during storage. Losses of mercury occur for all types of containers and even in highly acidic solutions (1, 2 ) . The loss of cadmium by adsorption takes place only on the walls of glass containers and at a pH greater than 7. Since polymer surfaces do not interact with cadmium in aqueous solution, sampling for cadmium would be better done in plastic containers. To prevent losses of cadmium by adsorption to the walls of glass containers, water samples should be acidified with nitric acid. If such pre-
cautions are taken, losses of cadmium from water samples during storage may be minimized.
ACKNOWLEDGMENT We wish to thank R. K. Tucker and R. A. Porter for helpful discussion. Received for review August 27, 1973. Accepted December 10, 1973. This work was supported in part by funds made available through the Research Council of the University of Idaho and Idaho State Office of Higher Education.
Fusion Methods for the Determination of Intractable Organic Compounds (Including Polymers) Sidney Siggia and David D. Schlueter D e p a r t m e n t of Chemistry, University of Massachusetts, A m h e r s t , M a s s . 0 1002
In organic analysis, we are often confronted with the quantitative determination of functional groups or compounds which react very slowly when solution methods are used. With solution methods, one is limited as to reagent concentration by the solubility of the reagent, reagentsample-solvent compatibility, and reaction temperature which is dictated by the reflux temperature of the solvent. Therefore, the application of fusion methods was obvious. With fusion, one uses pure reagent or thereabouts; the reaction is run at about melt temperature and can be raised to just below the thermal decomposition temperature of the sample components. Hence, for most organic compounds, fusion temperatures of up to 300-350 "C are common. Few compounds can remain unreacted under these conditions. Two factors control the resistance of a given chemical compound to chemical reaction. One is the inherent stability of the bonds of the molecules; the other is the steric configuration of the molecule. For example, amides, especially tertiary, are quite resistant to hydrolysis, yet caustic fusion brings about complete hydrolysis of these compounds in less than 30 minutes ( 1 ) . Polymeric esters, where the ester groups are pendant from a hydrocarbon polymer chain, are often very difficult to saponify in solution. Yet fusion with potassium hydroxide proceeds quite rapidly. For example poly(methy1 methacrylate) saponifies in solution only to the extent of 30% with 1N potassium hydroxide in amyl alcohol (bp 137 "C) after 90 hours a t reflux; by fusion with potassium hydroxide at 360 "C this compound is completely saponified in 0.5 hours ( 2 ) . Most chemists look on fusion reactions as being crude and somewhat akin to pyrolysis. This is not the case; fusions can be clean, quantitative, stoichiometric reactions as long as the fusion temperature is kept below the pyrolysis temperature of the compound under test. This is usually easy to accomplish since most fusion reactions take place between 200-350 "C and most decompositions in an oxygen-free system do not start until the temperature is above 400 "C. Fusion methods are far from being new. Alkali fusion reactions, dating as far back as 1840, have been used extensively in organic synthesis and degradation. Literature surveys of this important field have been published by ( 1 ) S . P. Frankoski and S. Siggia, Anal. Chem.. 44, 2078 (1972) (2) S . P. Frankoski and S Siggia, Anal. Chem.. 44, 507 (1972).
Weedon (3) and Nara ( 4 ) . Unfortunately, the analytical application of these reactions to organic analysis has not, until recently, been explored to any great extent. Fusion Reagents. These fall into four categories at present: bases, acids, reductants, and oxidants. In the case of the bases, sodium and potassium hydroxides have been the most widely used ( I , 2, 5-11) although lithium, rubidium, and cesium hydroxides have also been employed (11). Potassium hydroxide is generally preferred because of its lower melting point and the fact that more organic compounds dissolve in it. It has been found that a clean, homogeneous melt is almost always mandatory. The water (usually about 15% by weight) contained in commercial potassium hydroxide is essential in the fusion reagent, since many of the reactions are hydrolysis reactions. In addition, the water lowers the melting point to about 125 "C. Pure potassium hydroxide melts a t 360 "C. Only one acid fusion reagent has been successfully used to date (12). This is sodium hydrogen sulfate monohydrate (mp 58.5 "C). The water associated with the acid is needed with this reagent, as it was above, since hydrolysis is the reaction of interest. The amount of water as well as the melt temperature can be regulated by applying a VBCuum to the salt for a sufficient length of time. A number of hydrazines and hydrazides have been investigated as possible reducing agents in fusion reactions (13). Carbohydrazide was found to be the most promising because it is solid, easy to obtain, degrades slowly to evolve hydrogen and hydrazine, both of which effect re(3) E. C. L. Weedon. "Elucidation of Structures by Physical and Chemical Methods," Vol. XI, Part 2, in "Technique of Organic Chemistry," A. Weissberger, Ed., Interscience, New York, N.Y., 1963, Chapter X I I . pp 655-705. (4) K. Nara, Kagaku To Kogyo (Osaka), 43 ( l o ) , 539 (1969): ( 1 1 ) . 605; Chem. Abstr.. 73, 8 7 1 3 6 ~(1970). (5) S. Siggia, L. R. Whitlock, and J. C. Tao, Ana/. Chem., 41, 1387 (1969). (6) S. Siggia, L. R. Whitlock. and J. E. Smola, Ana/. Chem., 44, 532, (1972). ( 7 ) D. D. Schlueter and S. Siggia, in press. (8) M. G. Voronkov and V . T. Shemyatenkova, I z v . Akad. Nauk SSSR. Otd. Khim. Nauk, 1961, 178; C . 6. Trans. 164-5: Chem. Absfr.. 5 5 , 16285b (1961). (9) J. H. Wettersand R. C. Smith, Anal. Chem., 41, 379 (1969). (10) C. L. Hansonand R. C. Smith.Ana/. Chem., 44, 1571 (1972). (11) T. Kakagawa, K. Miyajima, and T. Uno, J . Gas Chromatogr.. 6 , 292 (1968). (12) Work in progress in our laboratory. (13) P. C. Rahn and S. Siggia. Ana/. Chem., 45, 2336 (1973). A N A L Y T I C A L C H E M I S T R Y , V O L . 46, NO. 6, M A Y 1974
* 773