in 10s is theoretically tolerable. Even though benzene is a stronger Raman scatterer than most organic molecules, dilute samples can certainly be run. Spectrum Sharpening. The quantity @ is proportional to the Raman cross section, g ~ and , the phonon lifetime, T ~ Since . the sharp features in a Raman spectrum are the features with large phonon lifetimes, the enhanced phonon Raman effect will enhance the sharp features. Thus, perhaps with this method, sharp features can be extracted from the broad bands in large organic molecules. Experimental Simplicity. For dilute samples, the momentum matching condition of Figure 2 is dependent on the index of refraction of the solvent. Then the angle Os is nearly independent of the solute and different biochemical samples can be run in succession with small changes in the experimental setup. This may allow computer solutions of the value of x (mirror position) as a function of frequency instead of a feedback loop for x (see Figure 5). Resonant Raman Enhancement. The fact the primary laser beam frequency in Figure 1 originates from a dye laser opens up the possibility of tuning the primary laser frequency through the visible spectrum. There is experimental Raman work indicating that when the laser excit-
ing frequency is tuned to the frequency of a chromophoric absorption band, the chromophoric phonon modes are greatly enhanced (2, 11). For example, in nitrobenzene, the NO2 phonon is strongly enhanced as the laser frequency shifts to the blue (11). In the classical Raman effect, fluorescence becomes a problem as one approaches absorption bands. As mentioned earlier, fluorescence will not be a problem on the anti-Stokes side of the laser frequency. This resonant enhancement could be a means of compensating for the weaker scattering cross sections of resonance for biochemical samples. I t has been shown in this paper that with modified experimental techniques, the acquisition of Raman data of large molecules is quite possible.
ACKNOWLEDGMENT The authors have profited from valuable discussions with P. Frazier Williams, Josi! Salzberg, and Herndon Williams. Received for review September 18, 1972. Accepted April 9, 1973.
Quantitative Analysis of Aqueous Nitrite/Nitrate Solutions by Infrared Internal Reflectance Spectrometry R. T. Yang and M. J. D. Low Department of Chemistry, N e w York University, N e w York, N. Y . 10003
Infrared spectra of aqueous NO2- and NOS- solutions, and of binary mixtures, were recorded using internal reflection techniques and a Fourier transform spectrometer. NO2- could be detected at a concentration of 0.02M, and NO3- at 0.001M. However, NOS- also became adsorbed on the Ge prism. After making corrections for NO3- adsorption and for overlapping of NOz- and NOSbands, it was possible to analyze binary mixtures with concentrations as low as [N03-] = 0.03M and [NOn-] = 0.05M. By adjusting the optics so that the usually totally absorbing water stretching band had a transmittance of about 50%, it was possible to observe N-H and C-H stretching bands of solutes. Dissolved materials cause the water structure and consequently the water absorption to change, causing distortions of the background of ratioed spectra of aqueous solutions. These distortions may influence qualitative and quantitative observations.
The utility of infrared internal reflection spectrometry organic solvent a t the liquid-solid interface ( I ) led us to attempt to study adsorption from aqueous solution. using a Ge IRS prism as adsorbent as well as for optical purposes. NO3- became adsorbed on Ge. prompting us to examine the quantitative aspects of the infrared analysis of aqueous solutions of NOa-, NOz-, and of NOs--NOzmixtures.
(IRS)in providing information about adsorption from
( 1 ) R T Yang M J D Low G L Haller and J Fenn
terface Sci 2014
in
J Colioid In-
press
A N A L Y T I C A L CHEMISTRY, VOL. 45,
IRS analysis of aqueous NaNO3 solutions has been described by Katlafsky and Keller (2), who found a linear variation, [NOS-] = aA(N03-), of the 1335-cm-l NO3absorbance over the 5-2070 concentration range. A ( N 0 3 - ) is the NO3- absorbance. Wilhite and Ellis ( 3 ) ,using IRS, reported a similar linear variation for the 0.1-1M NO3range, but another linear relation, [N03-] = b + cA(N03-), for the 0.01-0.1M Nos- range. Ahlijah and Mooney ( 4 ) used IRS to analyze aqueous NaNO3-NaN02 solutions; their calibration plots were linear but their graphs ( 5 ) show them to be [N03-] = d + eA(N03-) and [NOz-] = f g A ( N 0 y ) . Values of d and f are near 0.08 and 0.1 absorbance. As our observations differ somewhat from those made earlier, and the method is of potential utility, we describe our IRS study of aqueous NO3+, NO3-, and K 0 3 - - N 0 2 - solutions.
+
EXPERIMENTAL A trapezoidal Ge single-pass plate, 52 X 20 X 1 mm, was used in conjunction with a suitable variable-angle attachment described elsewhere ( 1 ) . Spectra having a constant resolution of 8 cm-1 were recorded with a modified Digilab Model FTS-14 Fourier transform spectrometer (6). All single-beam spectra were obtained by summing 400 interferograms. After the cell containing ( 2 ) B. Katlafsky and R . E. Keller, Ana/. Chem.. 35, 1665 (1963). (3) R. N. Wilhite and R. F. Ellis. Appi. Spectrosc.. 17, 168 (1963). (4) G . E. 6 . Y . Ahlijah and E. F. Mooney. Spectrochim. Acta.. 25A, 619 119f9) ( 5 ) Ref. 4 . figure 5. (6) M. J. D . Low, A. J. Goodsel. and H . Mark, in "Molecular Spectroscopy 1971." P. Hepple, E d . , Proc. 5th Conf. Molecular Spectroscopy, Brighton. England, 1971; Institute of Petroleum, London, 1972. \
- - - /
p p 383ff.
NO. 12,
OCTOBER 1973
A
D
A
i
Y
3000
CKi’
, . . , . . . . , . . , . ,. . . . , . . . , , . . . . , . ’ ~ 3000
lob0
2000
2000
1000 Cni
Figure 1. Single-beam spectra (A) Dry Ge plate in empty cell; (B) water; ( C ) aqueous solution 0.2M in NaN03 and 0.2M in NaN02: ( D ) aqueous solution 0.03M in NaNOr and 0.05M in NaN03
Figure 2. Ratioed spectra Each spectrum shown was obtained by ratioing two of the spectra of Figure 1. (A) water (B, Figure l / A , Figure 1 ) : (E) 0.2M NaN02 0.2M NaN03 solution ( C , Figure 1 / A , Figure 1); ( C ) 0 . 2 M NaN02 0.2M NaN03 solution ( C , Figure 1 / B , Figure 1 ) ; (D) 0 . 0 3 M NaN02 0.05M NaN03 solution ( D , Figure 1/B. Figure 1 )
+ + +
the Ge plate was installed, the sampling compartment was flushed with dry air. The cell remained in place throughout a series of measurements so that its transmittance remained constant; small changes in cell position can cause transmittance changes which can affect the results obtained when two single-beam spectra are ratioed. Also, the measured band intensities are dependent on factors such as the angle of incidence, beam convergence, detailed crystal geometry, and crystal index, which remained constant because a single experimental setup was used; band intensity values would change if a different setup were to be used. Pure water or solutions of various NHdN02, “4N03, N a S O j , NaSOn, or NaN02 concentrations. prepared from distilled water and reagent-grade chemicals, were placed into the cell or were withdrawn from it by means of syringes and Teflon (DuPont) tubing was permanently attached to the cell.
RESULTS AND DISCUSSION Only single-beam spectra were measured. These were stored by the instrument’s disk memory, retrieved, and ratioed by computation as required, using the instrument’s digital computer. Some examples of single-beam spectra are shown in Figure 1. The equivalent penetration depth ( 7 ) , computed by taking the molar extinction coefficient of liquid water to be 81 a t 3450 cm-I ( 8 ) , was 3.1 p a t 3450 cm-I and about 8 p a t 1600 cm-I. Spectrum A is the “background” of the Ge plate in air. There were prominent COz bands, and some water vapor bands in the 1900-1400 cm-1 region, caused by residual water vapor in the spectrometer. When water was in contact with the Ge plate, the prominent OH stretching and deformation bands were superimposed on the background (B, Figure 1).The single-beam spectra of dilute solutions (C, D, Figure 1) are similar to the spectrum of water; the bands of the solute are not prominent. The instrument function was removed when two such single-beam spectra were ratioed. If the empty cell was used as reference, the ratioed spectra had the appearance of spectra obtained by the usual double-beam operation, e . g . , A, B, Figure 2. When water was used as the reference, the ratioed spectra had the appearance of double-beam, solvent-compensated spectra (C, D, Figure 2 ) . The slight COz bands which appeared in some spectra were brought about by changes in ( 7 ) N. J. Harrick. “Internal Reflection Spectroscopy,” Interscience, New York, N . Y . , 1967. (8) W K . Thompson, Trans. Faraday Soc.. 61, 2635 (1965).
Figure 3.
1500 1300 Scale-expanded spectra
1100 cn
+
(A) 100% line see text ( B ) 0 2M NaNO? 0 2M NaN03 solution scale-expanded segment of spectrum B of Figure 2 ( C ) 0 03M NaN02 + 0 05M NaN03 solution scale-expanded segment of spectrum D of Figure
2
the efficiency of the air dryer. The noise above 3000 cm-1 and below 1000 cm-1 in spectra C and D of Figure 2 was the result of ratioing spectra in regions where the transmittance was very low. In spectrum C, the negative OH stretching band, the negative OH deformation band, and the increase in absorption below 1000 c m - l , which are barely detectable in spectrum D, are not artifacts. These changes were caused by variations in the absorptivity of the water present in the solution, the water structure being affected by the solutes. The absorptions of NOS- and NOz-, barely detectable in the single-beam spectra, became more recognizable in the ratioed spectra (A, B, Figure 2); computer-produced scale expansion then made the bands usable. Examples are shown in Figure 3. Trace A of that figure was obtained by recording two single-beam spectra of water and ratioing them against one another, in order to obtain an estimate of the noise level; it can be taken as a “100% line.” Spectrum B shows the NOs- and N0z- bands. Spectrum
A N A L Y T I C A L C H E M I S T R Y , VOL. 4 5 , NO. 12, OCTOBER 1973
2015
0 0.0005
v
1500
II
cm-'
Figure 5. Calibration curves
C is a highly scale-expanded segment of spectrum D of Figure 2, plotted to the same scale as spectrum C of Figure 3; the noise level is consequently higher. The ratioing and scale-expansion techniques outlined were used to examine single and binary solutions of various concentrations. Some segments of spectra of NH4N02 solutions are shown in Figure 4. Similar results were obtained with NaNO2 solutions, the intensity of the 1237cm- 1 NOz- band decreasing progressively with decreasing N02- concentration. A reasonably linear relation between the absorbance of the 1237-cm-1 band and NO2- concentration was obtained (Figure 5). The lowest concentration yielding a usable spectrum was 0.02M. Similar measurements were carried out with NO3- solutions. Some results obtained with Na?J03 are shown in Figure 6. The intensity of the NO3- band decreased progressively as the YO3- decreased, but remained constant a t concentrations less than 0.001M (Figure 6A). With relatively concentrated solutions, the NO3- band was broad and had a maximum a t 1352 cm-I and a shoulder near 1400 cm-I (Figure 6 B ) . (The weak symmetric stretching band of NO3- could also be observed near 1050 cm-1 with 1M N o s - solutions.) When the NO3- concentration declined, the 1400-cm-1 absorption diminished and the maximum of the residual, sharp band shifted to 1350 cm-I. The N o s - band of the more concentrated solutions is more like that of solid NaN03, for which the asymmetric stretching band is a t 1374 cm-I (9). (9) M Anbar M Halman andS Pinchas J Chem SOC 1960, 1242
2016
Figure 6. Spectra of NaN03 solutions The numbers next to each trace indicate [ N 0 3 - ] in M All spectra are scale-expanded Part A shows segments taken from a series of spectra all spectra of the series being scale-expanded by the same factor band intensity is consequently proportional to concentration The same segments are shown again in Part B In order to make changes in band contour more detectable only the absorption in the 1550-1000 c m - ' range was expanded For each segment the lowest absorption within that range was placed at the bottom of the chart and the highest absorption at the top of the chart so that the peak heights are the same The expansion factor is different for each segment and is supplied by the computer
The sharp residual NO3- band, which could not be diminished by extensive flushing with water but which could be removed by treating the Ge plate with hot chromic acid cleaning solution. is attributed to an NO3species fairly strongly bound to the Ge surface. Using the absorbance of the residual peak, and assuming that the species causing the band was N o s - , and that the area of the Ge plate was 20 cm2, leads to an estimate of one adsorbed unit per 5-10 A of surface. The total NO3- absorbance at 1352 cm-I could be corrected by subtracting the absorbance of the residual band of the adsorbed material. When this was done, the absorbance of NO3- in solution was found. and a reasonably linear relation between the corrected absorbance values and NO3- concentration was obtained (Figure 5). From that plot, the absorbance of the residual band (0.0111) is equivalent to the absorbance of -0.05M N o s - solution. Using the extinction coefficient of NO3- for the adsorbed species and estimating an equivalent penetration depth of about 7 p a t 1350 cm-l leads to an estimate that the absorbance of the adsorbed species was equivalent to the absorbance of a solution of about 0.06M N03-, in fair agreement with the value obtained from the N o s - calibration curve. However, the calibration plots of Figure 5 cannot be used directly for estimating the concentrations of NosNO2- mixtures. The 1237-cm-l NO2- band (asymmetric stretching) has a shoulder at 1340 cm-I (symmetric stretching) which, as shown clearly by Figure 7, overlapped the 1352-cm-l NO3- band. The latter does not cause the absorbance of the 1237-cm-l NO2- band to change significantly. The ratio of the absorbance at the 1340-cm-1 shoulder and that of the 1237-cm-l band is about 4:21. Taking the ratio of the absorbances of the NOz- asymmetric and symmetric stretching bands to be constant then gives [absorbance a t 1352 cm-I due to NO3-] = [total absorbance at 1352 cm-l] - [absorbance a t 1237 cm-l/5.25]. Some results obtained with dilute solutions are summarized in Table I. The correction for NO3- has been made. The agreement between actual and found values is not
A N A L Y T I C A L CHEMISTRY, VOL. 45, NO. 12, OCTOBER 1973
Table I . Analysis of NaN03-NaNO2 Solutions M NO*M NOSM NOSactual
actual
found
M N03found"
0.20
0.20
0.20
0.21
0.050 0.050
0.050
0.059
0.050
0.030
0.064 0.14 0.044
0.058 0.18
0.15
0.150 0.050
0.06 0.033
Corrected for adsorption and band overlapping.
Table II. Absorptions of N a N 0 2 and N a N 0 3 I RS
This work 1352 1047 1340 1237
Katlafsky Ahlijah and and Keller Mooney (41 (2) 1335
1225
-
Transmittance (77)
(9. 72)
Assignment
1333
1370
1385
1045
1050
1049
asym. stretch sym. stretch NOz- asym. stretch sym. stretch
1316 1220
1335 1240
unreasonable. If the correction for the band overlapping had not been made, the NO3- concentration values would have been too high by about 115 of the concentration of NO*-. It is pertinent to note that the lowest concentrations reported earlier ( 4 ) for a binary mixture were [N02-] = 0.7M, [NO3-] = 0.5M, 14 and 17 times as great as the lowest values in Table I. There are considerable differences in the frequencies of the NOa- and X 0 2 - absorptions which have been reported; a summary is given in Table 11. The dispersion of the index of refraction in the infrared region is known (7); i . e . , in the vicinity of an absorption band, the index of refraction first decreases, then increases sharply, and decreases again to approach the normal value after the absorption has been passed. Consequently, with IRS, the penetration depth reaches a maximum when the index of refraction reaches a maximum and absorption increases. When the sine of the incident angle is very close to the ratio of the refractive indices, the absorption band maximum is distorted toward lower frequency. In the present experiments, Ge and an average angle of incidence more than 60" were employed. The sine of the incident angle and the refractive index ratio of the Ge and liquid are far apart (0.87 and 0.33). The effect of the dispersion of the refractive index of the aqueous solution on the absorption band positions was therefore negligible. However, Katlafsky and Keller ( 2 ) used an Irtran-2 plate with an incident angle of 40". The critical angle for the system ZnS(Irtran-2)-€~20in the infrared range is 37.5%. Consequently, the incident angle was quite close to the critical angle, and any dispersion of the refractive index would cause a shift of absorption band maxima. Similarly, Ahlijah and Mooney ( 4 , 10) used a KRS-5 plate and stated that the incident angle was 30". However, the critical angle for the KRS-5-HzO system is 33.7", the refractive index of KRS-5 and H2O being taken as 2.4 (7) and 1.33, respectively. Their measurements must have been made with the incident angle clos'e to the critical angle, so that, again, band distortions occurred. The low values of band positions reported by Katlafsky and Keller and by Ahlijah and Mooney can be attributed to such band distortions. (10) G. E. B. Y . Ahlijah. and E. F. Mooney, Spectrochim. A c t a . 22A, 547 (1966). (11) L. V. Volod'ko and L. T. Huoah, Zh. Prikl. Spektrosk.. 9. 644 (1968). (12) R. E. Weston, J r . . and T. F. Brodasky, J . Chem. Phys.. 27, 683 (1957).
Figure 7. Spectra of N o s - and NO2
I..,....I....,....I..._, . . . , .
3000
2000
I
1000 Cni'
Figure 8. Spectra of H 2 0 and N H 4 N 0 3 solutions (A] Water; (B) 2M N H ~ N O Jsolution. The reference used for the ratioed spectra A and B was the single-beam spectrum of the dry Ge. ( C ) 2M N H ~ N O Jsolution, the ratio of spectra B and A . (D) 1 M N H a N 0 3 solution. the ratio of a spectrum (of 1 M NHaN03 solution similar to spectrum C) and sDectrum A
300
3000
2700 cn
Figure 9. Spectra of N H d + . Scale-expanded segments of spec-
tra C and D of Figure 8
It was of interest to determine how well spectral features of dissolved materials could be discerned in the vicinity of the intense water stretching band. Consequently, the incident angle was adjusted to decrease the penetration depth to about 0.8 p a t 3400 cm-I. The spectra of Figure 8 were then recorded. The water stretching band (A, Figure 8), obtained a t about 50% maximum transmittance, has the same contour as the band observed by transmittance measurements by
A N A L Y T I C A L C H E M I S T R Y , V O L . 45, N O . 12, OCTOBER 1 9 7 3
2017
Thompson et a2. (13). Two broad absorptions due to NH4+ (8) appear on the low-wavenumber side of the OH band in the 3100-2600 cm-1 region when NH4N03 is dissolved in the water (B, Figure 8), and these become more prominent when the spectra are solvent-compensated (C, D, Figure 8) and then scale-expanded (Figure 9). There is, however, some distortion of the 3 0 5 0 - ~ m -band ~ because of the negative OH band. The distortion decreases as the solute concentration is decreased, but this also decreases the solute band. Consequently, as the liquid layer is necessarily thin as well as dilute, the signal-to-noise ratio decreases. With the present technique, the lowest concentrations a t which N-H and C-H bands (similar experiments were made with aqueous methanol, Na citrate, etc.) were observed were near 0.1M. (13) W. K. Thompson, W. A . Senior, and 6. A . Pethica, Nature (London). 211, 1086 (1966).
The results thus show that it is feasible to obtain spectral information near or even at the position of one of the strong water absorptions. I t must be kept in mind, however, that the presence of solutes causes the water bands to change in absorption, so that distortions result. Such distortions may also affect quantitative results by causing base lines to fluctuate. The distortions decrease in intensity with decreasing solute concentration. However, because their intensities and frequencies are a function of the nature of the solute, they cannot be precisely compensated.
Received for review March 1, 1973. Accepted May 10, 1973. Support by Contract No. N00014-67-A:0467-0023 from ONR is gratefully acknowledged.
Spectrophotometric Determination of Total Amounts of Polythionates (Tetra-, Penta-, and Hexathionate) in Mixtures with Thiosulfate and Sulfite Tomozo K o h a n d K a z u o Taniguchi D e p a r t m e n t of Chemistry. Faculty of S c i e n c e , Tokai University, Hiratsuka-shi. Kanagawa. Japan
A sensitive determination of total amounts of tetra-, penta-, and hexathionate in mixtures with thiosulfate and sulfite is described. The method is based on the formation of thiosulfate equivalent to the polythionates, and on the spectrophotometric measurement of an excess of iodine for the thiosulfate formed. The factors affecting cyanolysis of pentathionate were examined, including a brief examination of cyanolysis of tetra- and hexathionate. The method could be successfully applied to the determination of these sulfur compounds mixed in various ratios with a deviation of = k O . O l umole.
In recent years, there has been a growing interest in the chemistry of polythionates, thiosulfate, and sulfite, and sensitive and accurate determination of these sulfur compounds in mixtures has become increasingly desirable. In previous papers (1-5), photometric methods for the determination of tetra-, penta-, and hexathionate, based on the formation of thiocyanate from polythionates, have been described. It has been possible to increase the sensitivity by approximately sixty-fold for the determination of polythionates by extracting an ion-pair between the cation of methylene blue and the anion of thiocyanate formed from the polythionates with an organic solvent (6). A method (7) was devised by the authors for the determination of polythionates in mixtures with each other, based (1) (2) (3) (4) (5) (6) (7)
2018
0. A . Nietzel and M. A . DeSesa. Anal. Chem.. 27, 1839 (1955). P. J. Urban, Z. Anal. Chem.. 180, 110 (1961). T. Koh, Bull. Chem. SOC. Jap.. 38, 1510 (1965). T. Koh and I. Iwasaki, Bull. Chem. SOC.Jap., 38, 2135 (1965). T. Koh and I . Iwasaki, Buil. Chem. SOC.Jap., 39,352 (1966). T. Koh. N. Saito, and I. Iwasaki, Ana!. Chim. Acta. 61, 449 (1972). T. Koh and I . Iwasaki, Bull. Chem. Soc. Jap.. 39, 703 (1966)
on the determination of different amounts of thiocyanate formed by the cyanolysis of polythionates in the presence and absence of cupric ions. However, no similar consideration has been given to the determination of total amounts of polythionates in mixtures with one another. Polythionates react with cyanide according to the following Equation:
S,06*-
+ (X - 1)CN- + HZO SO4*- + 2HCN + (X - 3)SCN- + S203'-
(1)
where X = 4 , 5 , and 6. The present paper is concerned with the determination of total amounts of polythionates when tetra-, penta-, and hexathionate are simultaneously present. The method is based on the formation of thiosulfate equivalent to the polythionates, and on the spectrophotometric measurement of an excess of iodine for the thiosulfate formed. A procedure for the analysis of mixtures of polythionates, thiosulfate, and sulfite is also described.
EXPERIMENTAL Apparatus. Spectrophotometric measurements were made with a Shimazu Model QV-50 spectrophotometer with 10-mm quartz cells. pH measurements were made with a Hitachi-Horiba Model M-5 pH meter. Temperatures were regulated by a Taiyo Model M-1 thermostat. Reagents. All other chemicals used, besides polythionates, were of an analytical grade and were used without further purification. Polythionates. Potassium tetrathionate was prepared as described by Stamm e t al. ( 8 ) , and potassium pentathionate and potassium hexathionate as described by Goehring and Feldmann (8) H Stamm, M . Goehring, and U . Feldmann, 2. Anorg Chem 250, 226 (1942).
A N A L Y T I C A L C H E M I S T R Y , VOL. 45, NO. 12, OCTOBER 1973
Allgem.
