and 7 about 2 sec or less). Then, from the pseudo-firstorder constant (0,’i sec-’ or more), the lower limit of the second-order rate constant of 2 x IO3 M - l sec-l is obtained. T h e second order rate constant 2 X lo4 (using Equation 3), 1 x 103 from 2 value-ha plot, and 2 x IO3 M - l sec-l from reversal of scanning, all confirm t h a t Equation 3 can be used for reactions described by Equations 4 and 7 . T h e importance of such a study is apparent: the rate of protonation of anion or radical anions and similar fast reactions can be studied. The pseudo-first-order condition for the reaction (Equation 4) was approximately maintained by considering ( a t 2 V/min) the peak current (about 5 FA) and the volume of cm3, plus t h a t of the hanging mercury drop (about 4 X the diffusion layer around the electrode. The concentration of the free radical generated, thus, comes out to be less mole/liter. (The concentration of 4-NBC1 than 2 X was 5 x 10-4M.) The experimental details have already been described (2). The concentration of pyridinium salt was 1 X 10-4M.
ACKNOWLEDGMENT T h e author gratefully acknowledges the facility provided by the Chemistry Department, State University of New York, Stony Brook, NY.
LITERATURE CITED (1) R . S. Nicholson and I. Shain, Anal. Chem., 36,706 (1964). (2) M. Mohammad and E. M. Kosower, J. Am. Chem. Soc., 93,2709, 2713 (1971). (3) E. M. Kosower, “An Introduction to Physical Organic Chemistry,” Sec. 2.8, Wiley, New York, NY, 1968. (4) M. Mohammad and E. M. Kosower, J. Phys. Chem., 74, 1153 (1970). (5) M. Mohammad, J. Hajdu, and E. M. Kosower, J. Am. Chem. Soc., 93, 1792 (1971). (6) M. Mohammad, unpublished results.
Mahboob Mohammad Department of Chemistry University of Islamabad Islamabad, Pakistan
RECEIVEDfor review August 12,1974. Accepted December 20, 1974.
Preparation and Properties of a Strontium-Selective Electrode Sir: A simple, rapid method for trace strontium determination was required to support process development for our nuclear waste management program. Levins ( I , 2) in his report describing a barium-selective electrode, mentions an analogous strontium-selective electrode. The preparation and properties of the strontium-selective electrode suggested by Levins were investigated. The electrode, a liquid-membrane type, is based on a complex of strontium with polyethylene glycol, which acts as a neutral carrier. T h e electrode is selective a t 0.1M concentrations for strontium in the presence of calcium and other divalent ions (with the exception of barium and mercury), and in the presence of alkali metals, except cesium. T h e electrode is limited in determining trace strontium in the presence of other cations, however, because of diminished selectivity inherent a t lower concentrations. EXPERIMENTAL R e a g e n t s . Deionized water and reagent-grade chemicals were used to prepare solutions of strontium and other salts. “Igepal” CO-880(Trademark, General Aniline and Film Corporation). which is nonylphenoxypoly(ethyleneoxy)ethanol, was purchased from Varian Instrument Division a n d used without further purification. T h e 4-ethylnitrobenzene ( E N B ) , from Aldrich Chemical Co., Inc., was purified by vacuum distillation. T h e neutral carrier complex salt was prepared as follows: 1.5 g (1 mmole) of “Igepal” CO-880was dissolved in 100 ml of H20; 5 ml of p H 6 buffer (60 ml glacial HC2H302, 270 g NaC2H302.3H20 per liter) were added, followed by 1 ml of 1M Sr(NO& or SrC12. T h e insoluble salt was precipitated by adding dropwise with stirring 10 ml of 0.5Msodium tetraphenylboron. After aging a t room temperature overnight, t h e clumps of white precipitate were washed with water, filtered onto a sintered glass filter of medium porosity, a n d subsequently vacuum-dried overnight. Without t h e p H adjustment, a finely divided unfilterable product was formed. Product a t room temperature was stable for severvacuum-dried over P 2 0 ~ al months; product vacuum-dried a t 50 “ C tended t o decompose within a few weeks. as evidenced by a benzene aroma. T h e strontium product obtained was assumed to be analogous to t h e barium compound described by Levins ( 1 ) and hence t o have t h e formula Sr . “Igepal’‘ CO-880 2B(C&)4. A solution of 0.2 to 0.5 g of the strontium compound in 5 ml of E N R served as the organic “exchanger” in t h e liquid membrane electrode. T h e particular concentration did not affect t h e response of t h e electrode. T h e most concentrated solution was thixotropic, which necessitated occasional agitation of t h e electrode.
A p p a r a t u s a n d P r o c e d u r e . T h e strontium-selective electrode was assembled in an Orion (Orion Research) liquid membrane electrode body No. 920000. Orion porous membranes were soaked in petroleum ether t o remove t h e proprietary treatment and were air-dried before use. T h e organic “exchanger” in t h e electrode was t h e solution of strontium complex in E N B . T h e internal aqueous solution was 0.1M SrC12 saturated with AgC1. T h e reference electrode was a double junction reference electrode, Orion No. 900200, with the outer chamber filled with 3M lithium trichloracetate. T h e electrodes were placed in -25 ml solution t h a t was slowly stirred magnetically. Potential measurements were made a t room temperature (-23 “C) with a n Orion Model 801 Digital pH/mV Meter.
RESULTS AND DISCUSSION Electrode Response. The measured potential as a function of strontium ion activity and concentration is shown in Figure 1. At >10-5M, the response is -27 mV per decade of activity. From 10-5 to 10-6M, the response is smaller, yet sufficiently reproducible to enable semiquantitative determinations in this region. Several electrodes were assembled, with different Orion porous membranes and with organic solutions of different concentrations. All had Eo values within 10 mV of one another. The lifetime of an electrode was several weeks, with a change in Eo, usually to a more-negative potential reading, of perhaps 5 mV. Electrodes were stored upright in air. Better response was maintained if the electrodes were never rinsed with water, but rather were blotted dry after use. During the working day, the Eo varied less than 1 mV. In pure SrC12 solutions, uncertainty in the measured potential was -0.2 mV, which would correspond to an uncertainty of -2% in strontium activity (or concentration). In 1O-I to 10-3M solutions, stable readings were usually obtained within 1 minute. With more-dilute solutions, a 5 minute period was required. At 10-6M, sometimes as long as 15 minutes was necessary to obtain a stable reading. Stability of readings was better with slow, rather than rapid, stirring. Effect of Other Ions. Selectivity constants (Ks,M) of the electrode for strontium over other cations were determined by the conventional method (3)from potentials ( E s , and E M ) measured in the respective solutions of 0.1M ANALYTICAL CHEMISTRY, VOL. 47, NO. 6, M A Y 1 9 7 5
959
Table I. Selectivity Constants for the Strontium-Selective Electrode (0.1M Solutions) M z+
KsrM
(CH3)iN' CS'
> I x 103
ca2+
2 X 102
Zn2+ Ni2' CO!'
8X 2x 2 x 10-3 2x 5 X 10-4 1 X 100 3 x 102
K' N a'
Li' "4'
L IO-
0-6
Sr",
13-2
c- 3
10-4
H' Sr2+ Ba2'
- 1
mol-!-'
hlZ+
KSrM
2 x 10-3 2x 9 X IO"' 8
X
lo-'
8 X IO-'
Fez'
7 x io-' 7 x 10-4 4 x 10-3 2x
Mg2+ Mn2+
Fe3+ ~13+
Figure 1. Typical calibration curves for strontium-selective electrode in SrCI2 solution
*iITT
:-
Table 11. Strontium-Calcium Selectivity Constants at Different Concentrations
P1 !.
15
&
c
01
--
_'?-'hi SrCI,
1
f
a"
-5
4
*: Different
Electrodes
Concn o!
S r C l 2 or CaClZ, !I
KSrCa
10-1 10-2
2 x 10-3
i
G '5F
10-3
8X 5 x 10-2
10-4
3 x 10-1
Table 111. Sr2+Activity i n SrC12 Solutions Concn of
~ I
d 2
l 3
4
l 5
l 6
7
, 8
l 9
IO
l II
l
I
1
l
~
Figure 2. Effect of pH on response of strontium-selective electrode
SrC12 and 0.1M solutions of other metal chlorides, MCl,, and calculated by the Equation:
ANALYTICAL CHEMISTRY, VOL. 47,
NO. 6,
MAY 1975
(5) (1) (1) (3) (4)
Table I\'. Average AMz+ Activities i n 0.1M MC1 Solutions Ionic
asr and
960
Reference
0.266 X 10" 0.55 X lo-* 0.79 x 10-3 0.93 X 10-j 1.0 x 10-5
10" 10-2 10-3 10-4 10-5
PH
where z is the charge of the interfering cation M; and U M are the respective single ion activities of the cations in 0.1M chloride solutions, computed as described in the next section. T h e denominator, 27, is the experimental slope of the Nernst plot. T h e reciprocal of K s r may ~ be regarded as the number of times as much M as Sr t h a t will cause significant interference in the measurement. These measurements in separate solutions a t 0.1M provide a general guide to interferences from other cations. Considerable drift in the potential was encountered in most MC1, solutions; the potential reading after 3 to 5 minutes was used for the computations. Selectivity constants are given in Table I. Response to (CH3)4N+ and Cs+ probably is due to the insolubility of these tetraphenylboron salts. Hg2+ affected the Eo of the electrode to the extent t h a t several hours of soaking in 0.1M Sr2+ solution were required for recovery. Barium is preferred over strontium; the relative selectivity is similar to t h a t found by Levins (2) with the barium-selective electrode. Common anions, such as nitrate, sulfate, silicate, and chloride, do not interfere. The effect of concentration on selectivity constants is illustrated in Table 11, where Ksrca was determined from measurements in separate SrClz and CaC12 solutions from lo-* to 10-lM. Selectivity constants increase with decreasing concentration. It is evident from Equation 1 that, as the concentration approaches the limit of detection, Esr and E M will approach the same value, as will asr and U M . T h e selectivity
5rzL act,\ ib, M
SrC12, I
1213
Debye-
strength, z
If
Av I*
1
0.1 0.2 0.6
0.774
0.777
0.77
0.519 0.337
0.702 0.656
0.28 0.046
2
3
H&kellCi
Av yMz+
Av
aMz*
0.77 x lo-' 0.28 x lo-' 0.046
x lo-'
constant will then approach unity, i.e., there is no selectivity. Thus ion-selective electrodes are necessarily limited in trace analyses to solutions t h a t contain minimal concentrations of other ions; selectivities determined a t higher concentrations are misleading. T h e p H effect in lo-' and 10-3M SrC12 solutions was determined by additions of HCl or NaOH to adjust the p H of the solution. In Figure 2, the change in potential is given as a function of the measured pH. The useful p H range for the electrode is 4 to 10. In the 10-IM solutions, the potential drifts a t high pH, as indicated in Figure 2. Estimation of Single I o n Activities. Single ion activities for Sr2+ and Mz+, for the strontium calibration curve in Figure 1,and for calculations of the selectivity constants in Table I are summarized in Tables I11 and IV. For SrC12, the individual activity coefficients estimated by Kielland ( 4 ) were used from 10-5 to 10-2M. For 0.1M SrC12, the value derived by Bates et al. ( 5 ) from their hydration theory was used, with the assumption t h a t 0.1 molar and 0.1 molal concentrations could be equated.
For the MCl,, the mean ionic activity coefficients of each valence type were averaged from tabulated values (6). The average single ion activities in Table IV were estimated using the p H convention ( 7 ) . Although this convention is properly applicable only a t ionic strengths less than 0.1M, these activity coefficients are adequate for the present application.
LITERATURE CITED (1) (2) (3) (4)
R. J. Levins, Anal. Chem., 43, 1045 (1971). R. J. Levins, Anal. Chem., 44, 1544 (1972). D. Ammann, E. Pretsch, and W. Simon, Anal. Lett., 5 , 843 (1972). J. Kielland, J. Am. Chem. SOC.,59, 1675 (1937).
( 5 ) R. G. Bates, B. R. Staples, and R. A . Robinson, Anal. Chem., 42, 867 (1970). (6) R. A. Robinson and R. H. Stokes, "Electrolyte Solutions," 2nd ed (Revised), Butterworths, London, 1970. (7) R. G. Bates, Pure Appl. Chem., 36, 407 (1973).
Elizabeth W. Baumann E. I. du Pont de Nemours & Co. Savannah River Laboratory Aiken, SC 29801
RECEIVEDfor review October 7 , 1974. Accepted January 17, 1975. Work performed under USAEC Contract No. AT(0T-2)-1.
I AIDS FOR ANALYTICAL CHEMISTS Program ELAL: An Interactive Minicomputer Based Elemental Analysis of Low and Medium Resolution Mass Spectra Joseph E. E v a n s and N. B. Jurinski U S . Army Environmental Hygiene Agency, Aberdeen Proving Ground, Md. 210 10
In recent years, the popularity of the combination gas chromatograph-mass spectrometer (GC-MS) has increased markedly. This is due, in part, to the more stringent regulations set forth by OSHA and EPA requiring definitive identification of environmental contaminants. Another factor t h a t contributed to this increase was the advent of the minicomputer based dedicated data system. This allowed for an efficient means of collection, reduction, and analysis of large volumes of data, thus making feasible, for many laboratories, an instrument t h a t was once considered only a tool of research. This rapid increase, however, has created a deficiency in the number of qualified mass spectrometrists needed to operate the large number of new instruments and interpret the data they provide. Therefore, many laboratories rely heavily upon the use of libraries of reference spectra and computerized search routines for definitive identification of unknowns ( I ) . Although these are useful tools t h a t aid in identification, invalid conclusions can often result when these methods are used alone. The practice of confirming the identification of unknown spectra by analyzing a reference standard is an excellent but not always practical procedure. T h e diligent use of elemental analysis can help the typical support laboratory reach more reliable conclusions, while significantly reducing turnaround time. Although elemental analysis will not provide complete structural information, it does provide information as to the amounts of each element present. High resolution mass spectrometry is the most powerful method of determining elemental composition; however, the effects of the abundances of naturally occurring isotopes in low and medium resolution spectra can also lead to the elemental composition of many ions. Many mass spectrometrists, because of inexperience and the time involved, fail to compute an elemental analysis on each unknown analyzed in the laboratory. With a dedicated
minicomputer a t their disposal, this chore can be readily handled. It was for this reason, program ELAL was developed. The program will compute an elemental analysis on low or medium resolution spectra for C, N, H, C1, E r , F , S, Si, 0 , and P.
EXPERIMENTAL was designed t o be a callable subroutine t h a t can be included as a part of existing GC-MS software or as a separate program t h a t can be used when the d a t a system is not in use. T h e program asks for only five items of information: t h e mass and intensit y of t h e (A) peak and the intensities of t h e ( A l), (A 2). and t h e ( A 4) peaks. Either normalized or raw d a t a may be entered. Computation is based on a modification of t h e manual method of elemental analysis described by McLafferty ( 2 ) .With this method, elements are categorized as ( A ) . ( A + 11,or (A 2) elements depending upon the prevalence of isotopes having l or 2 additional mass units. T h u s , C1 with isotopes of 35 and 37 a m u is considered an ( A 2) element. T h e common ( A ) elements are H, F, and P; t h e ( A 1) elements are C and N;and t h e (A 2) elements are C1, Br, S, Si, and 0. ELAL
+
+
+
+
+
+
+
RESULTS AND DISCUSSION After the requested data are entered, ELAL normalizes 1) peak to them and then uses the abundance of the (A compute the number of carbons present. By means of a binomial expansion, the contribution of carbon to the ( A 2 ) peak is determined. The (A 2) elements are then computed in order of decreasing isotopic abundance. If sulfur and/or silicone are found to be present, ELAL corrects for their contribution to the abundance of the (A 1) peak 1) and (A 2) and then completely recalculates the (A elements. The contribution of oxygen to the ( A 1) peak is usually insignificant, so a correction need not be performed. Checkpoints are provided throughout the program to ensure that ELAL has not calculated an erroneous number of atoms for any one element. Chlorine and bromine are computed simultaneously by subtracting the abundances of
+
+
+
+
+ +
ANALYTICAL CHEMISTRY, VOL. 47, NO. 6, MAY 1975
+
961
