Figure 6. Analytical curves for molybdenum in steel. Analytical line, 313.3 nm. FRAG, 200 mL/min. CV, 4200 V
the five points on a sample. The precision was 2.1 to 10% for the peak-height method, and 1.0 to 12% for the integration method. No differences in precision for the peak-height and integration methods were found from the results shown in Table 11. Analytical Curves. The analytical curves for various elements were prepared with the peak-height and integration methods. Table I11 shows the concentration ranges of elements studied for the preparation of analytical curves. Figure 6 shows the analytical curves for molybdenum in steel. Each plot in Figure 6 represents the average of the values obtained a t the five points. The vertical lines on each plot represent the standard deviations. The analytical curves for molybdenum in steel were linear for both the peak-height and integration methods. The analytical curves for other elements also were linear for both methods. Detection Limits. Detection limits were determined by use of the samples containing the lowest concentrations of each element. The detection limit was defined as the concentration or weight of element which gives a signal equal to twice the standard deviation of the background level. The weight of the detection limit was calculated from the concentration of the detection limit and the weight of the sample sputtered with the laser shot. The weights of brass, steel, and aluminum alloy sputtered with one laser shot were about 1.0,O.g and 0.8 pg, respectively. In this study, the most sensitive analytical lines for most elements were not used for the determination of the detection limits. By multiplying the detection limits
obtained with less sensitive lines by the relative sensitive factors for each analytical line, the detection limits for the most sensitive lines were obtained for most elements. Table IV shows the detection limits obtained with the less sensitive lines and those calculated for the most sensitive lines. The detection limits for the most sensitive lines ranged from 1.9 ppm (1.9 X 10-l' g) of iron in brass to 32 ppm (2.6 X 10-l' g) of copper in aluminum alloy with the peak-height method. Those ranged from 6.4 ppm (6.4 X lo-'' g) of iron in brass to 120 ppm (9.6 X g) of copper in aluminum alloy with the integration method. As the time for measuring the integrated value was about one order of magnitude over that for measuring the peak value, the fluctuation of light from the hollow-cathode lamps affected the integrated value more than the peak value. Therefore, the detection limits obtained with the peak-height method were lower than those obtained with the integration method. Because the copper hollow-cathode lamp used picked up a ripple of 60 Hz derived from the power supply, the background level fluctuated. The detection limit for copper was poorer than that predicted. The detection limits for copper, iron, and manganese in aluminum alloy were poorer than those in brass or steel. As mentioned previously, the atomic vapor of aluminum disappeared most rapidly in the atomic vapors of copper, iron, and aluminum of the matrix elements. The aluminum of the matrix element is assumed to affect the behavior of elements coexisting in the aluminum alloy. Therefore, the detection limits for the elements in the aluminum alloy were poorer than those in brass or steel. The atomization-absorption cell used in this study does not always have the optimum geometry. If the diameter of the absorption cell is reduced, higher sensitivity might be brought about by the improved geometry, and the detection limits can be lowered considerably.
LITERATURE CITED (1) H. Massmann, "Flame Emission and Atomic Absorption Spectrometry", J. A. Dean and T. C. Rains, Ed., Marcel Dekker, New York, 1971, Vol. 2, Chapter 4. (2) V. G. Mossotti, K. Laqua, and W. D. Hagenah, Spectrochim. Acta, Part 8, 23, 197 (1967). (3) A. V. Karyakin and V. A. Kaigorodov, Zh.Anal. Khim., 23,930 (1968). (4) E. K. Vul'fson, A. V. Karyakin, and A. I. Shidlovskii, Zavod. Lab., 40, 945 (1974). (5) D. E. Osten and E. H. Piepmeier, Appl. Spectrosc., 27, 165 (1973). (6) J. P. Matousek and B. J. Orr, Spectrochim. Acta, Part 8,31, 475 (1976).
RECEIVED for review January 17,1977. Accepted May 12,1977.
Spectrophotometric Determination of Vitamin K Compounds J-C Vir6 and G. J. Patriarche Institut de Pharmacie, Universitg Libre de Bruxelles, Campus Plaine, 205/
1 Bruxelles, Belgium
Robert J. Nowak and Harry B. Mark, Jr." Department of Chemistry, University of Cincinnati, Cincinnati, Ohio 4522 1
A spectrophotometric method for the determination of Vitamin K (naphthoquinones) compounds has been developed. TItanium (111) is employed to reduce naphthoqulnone to its corresponding naphthohydroquinone which subsequently forms a highly colored complex with TI(1V). Complex formation as a function of time, sulfuric acid concentration, and Ti(II1) and Ti( I V ) concentration is reported. The method yields highly linear calibration data In the region of lo-' M.
Menadione (vitamin K3) and analogues (naphthoquinones)
are of considerable importance because of their role in blood coagulation processes. Thus, trace analytical techniques for these compounds are necessary for the study of their biological actions as well as in pharmacological and toxicological studies. Previously, these compounds were determined spectrophotometrically (1-3) and spot titanium(1V) tests for the identification of naphthoquinones as characteristic complexes had been reported ( 4 ) . Also, similar phenolic type compounds such as chromotropic acid (5), sodium 1,3-dihydroxybenzene3,5-disulfonate (6) and thymol (7) had been employed as spectrophotometric reagents for the determination of titanium. ANALYTICAL CHEMISTRY, VOL. 49, NO. 9, AUGUST 1977
1343
Thus, the purpose of this spectrophotometric study of the complexes found on the reaction of titanium(1V) and the naphthoquinones of the vitamin K series was to develop trace analytical techniques for the differentiation and quantitative analyses of these compounds, as reported below. The method is based on the formation of a colored complex which is obtained on the reaction of a solution containing an excess of a mixture of titanium(II1) and titanium(1V) in 14 M sulfuric acid. The presence of titanium(II1) in the reaction media is essential to reduce the naphthoquinones to the corresponding naphthohydroquinones which are the actual complexing agents which coordinate with excess titanium(1V). For example: OH
0
I
II
0
7CO
eo0 A NII
0.2
0
1,4-Naphthoquinone
Sodium 1,2-naphthoquinone4-sulfonate
2-Methyl-3-phytyl-l,4-naphthoquinone (Vitamin K , )
EXPERIMENTAL Apparatus and Reagents. All of the sulfuric acid solvent solutions were made by dilution of Merck P.A. concentrated sulfuric acid. The titanium(II1) stock solutions were 15% Ti2(S0J3in 23% H2S04by weight and the stock titanium(1V) was prepared by dissolving Tic&in water according to the method of Arthur and Donahue (8). Because of solubility problems, the solutions were prepared by stock 2-methyl-1,4-naphthoquinone dissolving 100 mg of the naphthoquinone in 100 mL of methanol as other compounds except sodium 1,2-naphthoquinone-4sulfonate, which was dissolved in water. It was necessary to prepare these solutions daily and maintain them in the dark because of decomposition problems. All spectra were taken with a Beckman Acta V spectrophotometer (chart speed 0.5 nm s-') at room temperature, and blanks of identical solvent composition were employed in the reference beam. RESULTS AND DISCUSSION For convenience, the spectral properties and optimization of the reaction variables are discussed for each naphthoquinone separately below. Menadione (Vitamin K3). A specific aliquot of the menadione stock solution is placed into a 25-mL volumetric flask and 0.5 mL of the Ti2(S04)3stock solution and 0.3 mL 1344
600
Flgure 1. The spectra of the different na htho uinones compiexed 1.: X MaTi4+: 2 with titanium. Medium: H2S04: 14 M; Ti': X M. (A) 2-Methyl-1,4-naphthoquinone: 1.1 X M. (B) 1,4-Naphthoquinone: 9.5 X M. (C) Sodium 1,e-naphthoM. (D) 2-Methyl-3-phytyi-l,4quinone-4-sulfonate: 1.2 X M naphthoquinone: 8.9 X
300
Menadione (Vitamin K,)
so0
Naphthohydroquinone, NHQ Ti(IV) (NHQ) ( 1 ) (colored)
Although Ti(1V) is a product of the reduction reaction of NQ to NHQ, the reaction goes to completion only when there is a large excess of titanium(1V) present. It is important to note that only the phenolic forms yield a colored complex with Ti(1V); quinonic forms are unreactive. The compounds examined in this study were: 0
1
400
0.6
OH
Naphthoquinone, NQ
1
ANALYTICAL CHEMISTRY, VOL. 49, NO. 9, AUGUST 1977
500
700
A,NM
Figure 2. The spectrum of menadione (2-methyl-l,4-naphthoquinone) in 14 M H2S04solution
of the TiC14stock solution (diluted twice with methanol to minimize the violence of the HC1 evolution when mixed with 14 M H2S04) are then added. After cooling, the volume of the solution is brought to 25 mL with 14 M H2S04and it is left in the dark for 30 min prior to use. Figure 1A shows a typical spectrum for menadione, using the above conditions, which exhibits a broad maximum at 645 nm, a peak a t 510 nm, and an acute peak a t 391 nm. I t is evident that there are actually two distinct complexes formed under these conditions, as the initial color of the reaction solution is blue which slowly changes to a deep blue-violet after standing 30 min in the dark. It is thought the peak at 510 nm would correspond to a red complex and the peak at 645 nm would represent some blue type complex. The ratio of the concentrations of these two complexes was found to be dependent on both the concentration of the Ti(1V) and the concentration of the H2S04. It was also found that the structural nature of the various naphthoquinones also has a strong influence on the color of the final solution. These points will be discussed later. Note that the spectrum of menadione in 14 M H2S04as shown in Figure 2 is significantly different and exhibits no peaks at 510 and 640 nm which indicates that there is a strong coordination interaction of Ti(1V) in the complex. A kinetic study of the evolution of the spectrum of the complex showed that the peak at 645 nm shows a steady decrease while concomitantly, the peak a t 510 nm increases and reaches a constant value after about 30 min. Effects of H2S04Concentration. It has been shown that the concentration of H2S04has a drastic effect on the nature of the formation of the Ti(1V) complex. Below 8 M HzS04, although it is possible to obtain a colored complex with Ti(1V) for sodium 1,2-dihydroxybenzene-3,5-disulfonateor chro-
/-
-
1.0
,A'
/ B
W
2 0
B
0'5
-
I
400
500
600
800 A.NM
700
Figure 3. Menadlone (Bmethyl-l,4-naphthoqulnone): study of spectra as a function of acidity. (A) H2S04, 15 M (B) 14 M, (C)13 M, (D) 12 M, (E) 1 1 M
0.0 00
1
I
1
1
I.o
2.0
30
4.0
CONCENTRATION, Mo%
Flgure 5. Absorbance as a functlon of the concentration of menadlone at 530 nm. Concentratlon of tltanlum: (A) TI3+ = 2.4 X M; Ti4+
=3x
=8X 5X
M.
(~ =p 1.6+ x io-2M; TI^+ = 2 x io-2M. (c)T?+ =1 X M. (D) Ti3+ = 4 X lom3M; TI4+ =
lom3M; TI
0.4
M
I
I
I
0
60
120
I
180
TIME, MINUTES
Flgure 6. Absorbance of menadione complex function of tlme at 530 nm
0.25
00 00
I
I
1
.008 010 CONCENTRATION, TOP:Ti3*, BOTTW: T i 4 *
I 016 020
Figure 4. Menadione: absorbance as a function of the concentratlon of tltanium. Concentratlon of menadlone: 2.3 X lom4M. (A) at the 510 nm peak. (B) at the 530 nm isosbestic polnt
motropic acid, no spectral evidence for the formation of a Ti(IV) complex with the naphthoquinone is observed. Figure 3 shows that the peak at 510 nm is observed only a t high H2S04concentrations; 15 M to 13 M in H2S04(curves A, B, and C, respectively) and that the magnitude of the 645 nm peak decreases markedly with decreasing acid concentration. The isosbestic point at 530 nm is of particular importance because at that wavelength the absorbance of the solution is essentially independent of the concentration of the HzS04 (there is a change of only 0.002 absorbance unit for a change of acid concentration of 13 M to 15 M). It was observed that for H2S04below 10 M, hydrolysis of Ti(1V) occurs which results in turbidity of the solution making optical measurements impossible. Effects of Concentration of Titanium Ions. It was found that the concentration of the complexing ion, Ti(IV), has a large influence on the degree of completion of the formation of the complexes as can be seen in Figure 4. The peak at 510 nm reaches a maximum value for Ti(IV)/Ti(III) concentration of 2.5 X M/2 X M, respectively, as shown in Figure 4. The peak at 645 nm, however, continues to increase with increasing concentrations of Ti(1V) and Ti(1II). The isosbestic point at 530 nm also increases in
(1.2X lob4M) as a
absorbance as the Ti(1V) and Ti(II1) concentrations increase, as shown in Figure 4. Figure 5 shows the variation of absorbance at 530 nm (isosbestic point) as a function of menadione concentration for five different values of Ti(IV)/Ti(III) concentrations. It is evident from this study that the Ti(1V) concentration must be about 2.0 X lo-' M and Ti(I1I) must be about 1.6 X lo-' M in order to obtain a linear calibration curve. Effect of Development Time. Figure 6 shows that the time required to reach a constant absorbance value at the isosbestic point at 530 nm is approximately 30 min (the optical density increases by 0.001 absorbance unit between 30 and 60 min). A calibration curve for menadione concentrations between 0 and 5.8 X M yields a straight line with a correlation coefficient of 0.9998, with e = f5150 when the absorbance was measured at 530 nm after 30 min. 1,4-Naphthoquinone. The complex formed by reaction with 1,6naphthoquinone,which differs from menadione only in the absence of a two position methyl group, shows a significantly different spectrum, as shown in Figure 1, curve B. The peak in the 645 nm region appears as only a small shoulder and there is a principal peak at 505 nm and smaller sharp peak at 375 nm. The optimum reaction parameters, [Ti(IV)],[Ti(III)],[H,SO4], and time were identical to those for menadione. The calibration curve for concentrations between 1.25 x and 7.6 X M 1,4-naphthoquinone as measured at the 505 nm peak was linear with a correlation coefficient of 0.99991 (e = f5590). Sodium 1,2-Naphthoquinone-4-sulfonate.With the carbonyl groups in the 1,2-position,the spectra of the resulting Ti(1V) complex are drastically different. An extremely broad ANALYTICAL CHEMISTRY, VOL.
49, NO. 9,AUGUST 1977
1345
single band with a maximum at 545 nm is observed with no evidence for a 645 nm peak. There is, in addition, an illdefined shoulder in the 385 to 430 nm region. As sodium 1,2-naphthoquinone-4-sulfonate in 14 M H&304in the absence of titanium shows a peak at 390, this shoulder could possibly indicate that the complexation reaction is not quite complete under these conditions. This complex is violet-blue in color. In contrast to the previous two compounds, the peak maximum at 545 nm does not change drastically at high acid strengths (13 to 15 M). Decomposition sets in above 15 M H2S04. As the effect of Ti(1V) and Ti(II1) concentrations was the same as that found for the previous two compounds, except that the absorbance was measured at 545 nm, the experimental conditions for the analysis of sodium 1,2-naphthoquinone4-sulfonate were identical as described above. A calibration curve for this compound from 7.7 X to 4.6 X was linear with a correlation coefficient of 0.999987 ( E = 4~4350) and, for the range 1.9 X to 2.3 X M, the correlation coefficient was 0.99994 (e = *4350). 2-Methyl-3-phytyl-1,4-naphthoquinone (Vitamin Kl). This molecule is different from the others because of the long side chain in the 3-position which perturbs complex formation. The spectrum (Figure 1,curve D) shows a single absorbance band with a rounded maximum between 640 and 655 mn. The two peaks at 510 and 390 nm have disappeared and the color of the complex is different. It is blue and corresponds to a single type of complex. The time needed to form the complex, 45 min, is longer, presumably because of steric hindrance. A calibration curve for this compound from 8.9 X lo4 to 8.9 X M was linear with a correlation coefficient of 0.99999 (e = *5730). Nature of the Complexes Formed. The structural nature of the complexes formed between titanium(1V) and these naphthoquinones is unknown although there is some indirect evidence that the less sterically hindered species, i.e., 1,4naphthoquinone, forms dimeric while the more sterically hindered 2-methyl-3-phytyl-1,4-naphthoquinone forms monomeric complexes. Menadione forms a complex which appears to be in equilibrium with its dimeric form. Evidence for these speculations is as follows: (i) All complexes were found to have 1:l stoichiometry (determined by Job curves).
(ii) The absorption spectrum of the menadione complex as a function of acidity indicates the presence of two species in equilibrium (monomer/dimer) as is evidenced by the isosbestic point at 530 nm. Further, lower HzS04 concentrations favor the species absorbing at 510 nm (dimer). (iii) The menadione complex shows a shift in the absorption spectrum as a function of time; the 510 peak increases (dimer) while the 645 nm peak decreases (monomer). (iv) The complex formed with the sterically hindered 2-methyl-3-phytyl-l,4-naphthoquinone exhibits a single peak a t 645 nm (monomer favored) while (v) the complex formed with 1,4-naphthoquinone exhibits one peak at 510 nm (dimer favored). It has been assumed that the presence of an aliphatic side chain (i.e., methyl, phytyl) alters the absorption spectra of these complexes through steric effects exclusively. Presumably all of these complexes are bound through Ti-042 bonds which are well documented. The concentration of H#04 determines the form of the Ti(1V) species which in turn affects the absorption spectrum of the complex (cf. Figure 3). If the concentration of HzS04 is too low, Ti(1V) hydrolyzes and complex formation with most naphthoquinones does not occur. The spectrum of the complex formed with sodium 1,2naphthoquinone-4-sulfonate is difficult to explain on the basis of these arguments because this ligand is quite different from the others studied.
LITERATURE CITED (1) E. E. Van Koetsveld, Red. Trav. Chim Pays-Bas, 69, 1217 (1950). (2) H. Wachsrnuth and R. Denissen, J. Pharm. Belg., 42, 173 (1960). (3) S. S. M. Hassan, M. M. Abd Ei Fatteh, and M. T. M. Zakl, Fresenius’ 2. Anal Chem., 275, 115 (1975). (4) I. M. Kolthoff and P. J. Eiving, “Treatise on Analytical Chemistry”, Part 11, Vol. 5, Interscience, New York, N.Y., 1961 p 1. (5) T. C . J. Ovenston, C. A. Parker, and C. G. Hatchard, Anal. Chim. Acta, 6, 7 (1952). (6) J. H. Yoe and A. R. Arrnstrong, Anal. Chem., 19, 100 (1947). (7) J. V. Griel and R. J. Robinson, Anal. Chem., 23, 1871 (1951). (8) P. Arthur and J. F. Donahue, Anal. Chem., 24, 1612 (1952).
RECEIVED for review March 10,1977. Accepted May 12,1977. Thanks are expressed to the Fonds National de la Recherche Scientifique (F.N.R.S. Belgium) and the National Science Foundation (Grant No. CHE76-04321) for support of this research.
Fourier Transform Infrared Spectrometric Determination of Gaseous Performic Acid P. D. Maker, Hiromi Nlki,” C. M. Savage, and L. P. Breitenbach Research Staff, Ford Motor Company, Dearborn, Michigan 48 12 1
Gaseous performic acid in the concentration range 0.01 to 1.0 Torr was analyzed using a Fourler infrared spectrometer system. Sample stabillty problems were overcome by the fast data acquisltion speed and high sensitlvity of the system. Sample purity problems were circumvented by taking full advantage of the system’s hlgh resolution and spectral subtraction Capabilities.
Definitive identification of performic acid (HC(=O)OOH) as a reaction product in the photo-oxidation of hydrocarbons 1346
ANALYTICAL CHEMISTRY, VOL. 49, NO. 9, AUGUST 1977
is crucial for the understanding of atmospheric chemistry ( I ) . However, because of gas handling problems, no reliable analytical technique has been available. According to Giguere and Almos (Z),gaseous performic acid decomposes spontaneously to formic acid even at 0 “C and is highly explosive in concentrated form. Nevertheless, they succeeded in recording its IR spectrum but with unknown sample pressure. Unfortunately, its major absorption bands are not readily distinguishable under low resolution from those of formic acid. The work reported here demonstrates that the IR Fourier spectroscopic method can overcome these difficulties. By combining the following essential capabilities, it becomes a
