from the top opening of the funnel. filtered through paper which has been dampened with benzene into a 100-ml. volumetric flask, and diluted to volume. The absorbance is measured spectrophotometrically a t 480 mH and the phthalate content of the aliquot taken is determined from a graph prepared with potassium biphthalate as standard. using a water solution of the standard and drying aliquots equivalent t o 0.5 to 4.0 mg. of phthalic anhydride.
Quantitative Procedure for Succinates. The procedure for succinates is identical to t h a t outlined for phthalates, except t h a t larger samples estimated to contain from 4 t o 25 mg. of succinic acid are used and the oven temperature is reduced to 135’ C. The red color is extracted with about 40 ml. of benzene, so t h a t the final dilution can be reduced to 50 ml., and the absorbance of these solutions determined a t 520 mp. A working curve can be prepared with succinic acid as standard, drying the aliquots a t temperatures below 100” C.
Table I.
Analysis of Typical Alkyd Resins
Phthalic Anhydride,
% Resin
Present (ultraFound violet) (colorimetric)
1 2 3 4 (phenol modified)
11.8 37.5 43.0 40,6
11.4,11.8 38.8,38.8 42.6,43.5 41.4,40.5
DISCUSSION
Accuracy of the method is not so good as some of the older methods for determining phthalic anhydride ( I ) , but its broader application and lack of interference from other acids and resins are unique and valuable. Duplicate determinations should always be made and all analyses and standardizations should utilize the same equipment. The yield
of colored condensation products varies with time and temperature of reaction. The 135” C. temperature established for succinates gives maximum yield of color. Highest yields in the phthalate method are attained a t temperatures above the prescribed 145” C., but the higher temperatures cause foaming of the reagent which begins to decompose under the attack of concentrated acid. This explains the need for good oven temperature control. The variations in points used to plot the working curve for phthalates (Figure 1) illustrate the anticipated precision. Some analytical results are shown in Table I. LITERATURE CITED
( 1 ) .4m. SOC. Testing Materials, “hSTM
Standards,” Designation D 1306 and D 1307, 1954. ( 2 ) Swann, M. H., Adams, M. L , Espoposito, G. G., ANAL.CHEW27, 1426 (1955).
RECEIVEDfor review December 18, 1956. .4ccepted A4pril27, 1957.
Spectro-Visual Method for Determining End Points Application to Titration of Soh ble Sulfate HARLEY H. BOVEE and REX J. ROBINSON Chemistry Department, University of Washington, Seattle 5, Wash.
F A new technique is described for detecting the end point in the titration of sulfate with standard barium chloride using the indicator tetrahydroxyquinone. The sample is illuminated with light transmitted by Wratten filter No. 45 and the end point is observed as a change from green to blue. The method is simple and the end point is sharp and reproducible. This principle of filter selection should b e applicable to other visual titrations.
T
volumetric determination of sulfate ion with barium chloride using the disodium salt of tetrahydroxyquinone as an internal indicator has been described by several investigators (4, 6-9, 11). The color change of the indicator is from yellow to orange-red but is not sufficiently sharp for accurate work, particularly by the inexperienced operator. I n an effort to improve and standardize the end point determination, Hallett and Kuipers (4) suggested comparison with light transmitted by two thicknesses of Wratten No. 21 red filter to indicate the near approach of HE
the end point. Ogg, Willits, and Cooper ( 7 ) carried this procedure one step further by taking as the end point the exact match of the sample solution with a standard polished glass filter having 37% spectral transmittance a t 550 mp, Both the sample and filter were lighted from below to assist the color matching. I n 1953, Storlazzi and Ransom ( I O ) noted that a sulfate titration viewed by light transmittance through a R r a t t e n S o . 45 blue filter gave a green to blue color shift a t the end point. The present work was directed towards investigation of this color change and establishment of an accurate and reproducible procedure for determining the end point. EXPERIMENTAL
Equipment and Reagents. To provide suitable underlighting, a light table was constructed from a metal box: two l b w a t t fluorescent lamps, and a translucent glass cover. The cover was masked with opaque paper except for two openings, to furnish light beams viewing the sample and reference solution simultaneously. No. 45 Wratten filters were mounted between glass
plates and secured over the light openings. 4 Fisher Electrophotometer was used for the photometric titrations. The spectral transmittance curves for the indicator were run on the Beckman DU spectrophotometer. The barium chloride solutions were approximately 0.004, 0.02, and 0.1N. These solutions were standardized against hydrazine sulfate, using the titration procedure described. The indicator reagent was tetrahydroxyquinone, supplied by W.D. & L. D. Betz Co.; it was used in the dry state. Selection of Filter. The visual color change in the titration solution may be understood from the spectral transmittance curves for the tetrahydroxyquinone indicator both before and after the end point shown in Figure 1. Similar spectral curves hal-e been obtained by Lee and others ( 5 ) and Walter (12). From curve 1 it is apparent that the indicator before the end point has a high rate of transmittance in the red portion of the spectrum extending well into the yellow before dropping to a minimum in the blue-green range a t about VOL. 29, NO. 9, SEPTEMBER 1957
1353
475 mp. This accounts for the yelloworange color observed for the indicator solution. At the end point the spectral transmittance shifts toward the red, as shown in curve 2, and the resulting solution appears orange-red to the eye. If an optical filter is now interposed in the light path, the combination of indicator and filter give spectral transmittance curves similar to the curves shown in Figure 2 for Wratten filter No. 45. These curves were calculated from the data used for Figure 1 and the spectral transmittance data for Wratten filters as given by Eastman Kodak ( I ) . The combined effect of the indicator and filter was taken as the product of their respective transmittances a t each wave length. Curve 2 in Figure 2 presents the effective light transmittance of the titration solution before the end point: it. peak value occurs at 510 mp in the green range. Curve 3, after the end point peaks a t 475 mp in the blue portion of the spectrum. Thus, by screening out the unwanted portion of the spectrum. it is possible to utilize only the wave lengths that give a color shift which is easily discernible. A large number of filters was screened by computing their effective transniittances in combination with the indicator curve. The most promising ones were checked visually by running samples. However, of all the filters considered. the most satisfactory for the visual titration n-as Wratten No. 45.
However, amounts of sulfate ion below approximately 0.5 mg. do not titrate stoichiometrically, but instead react very nearly the same as the blank. This anomaly has been noted by others (4, 7 , 8, 11) with the cutoff point variously estimated from 0.5 to 2 mg. I n Figure 3 the data have been plotted only to the cutoff points in order to compare better the lower limits with the different strengths of barium chloride. S o attempt has been made in the present work to explore the upper end of the sulfate range, the highest value titrated being 50 mg. of sulfate (2000 p.p.m.). Sheen and Kahler (9) titrated sulfate in concentrations up to 750 mg. per 25 ml. (30,000 p.p.m.) using tetrahydroxyquinone as the indicator.
i
i
:L 20 I
1 -!
400
500
WAVE L E N G T H
- UP
600
Figure 1. Transmittance curves for tetrahydroxyquinone 1. Before end point
2.
After end point
c
CALIBRATION CURVES
Csing the above titration procedure varying amounts of sulfate were titrated m-ith different concentrations of barium chloride. The data are given in Table I and the corresponding curve< are plotted in Figure 3. For values of sulfate above approximately 0.5 nig. the results fall on a straight line and follow a stoichiometric relationship.
1354
ANALYTlCAL CHEMISTRY
Additional information regarding the reactions of small amounts of sulfate have been obtained by photometric titrations. These were made using a Fisher Electrophotometer with a 525mp filter and a light intensity of D. This filter was selected because, as can be seen from Figure 1, the greatest change in density occurs a t this wave length. Walter ( l a ) also titrated photometrically a t about this wave length for the same reason. The series of curves shown in Figure 4 were obtained titrating different amounts of sulfate with 0.1N barium chloride solution. Inspection of the photometric curves reveals that for the blank there is a straight-line increase in density until all the indicator has reacted with barium ion. For sulfate values of 0.4 mg. and above, the density rises s l o ~ ~ - lowing y, to the precipitated barium sulfate until the equivalence point is reached, then rises rapidly until the indicator has h e n exhausted. The equivalence point is taken as the intersection of the horizontal and vertical portions of the curve. Between 0.0 and 0.4 mg. the character of the curves is in a state of transition. The titration curve for 0.1 mg. of sulfate (No. 2) is practically identical with the blank; apparently, the barium ions are reacting directly with the indicator and not a t all with the available sulfate. The curve for 0.3 mg. of sulfate ( S o . 3) appears to be going through a transition stage where the barium ions react with both the sulfate and indicator ions present a t the same time. At concentrations of 0.4 mg. of sulfate ( S o . 4) and above, the curves exhibit their expected form, showing that the barium sulfate is completely precipitated before any appreciable amount of barium tetrahydroxyquinone ic f o r n i d .
i u
Titration Procedure.
The sulfate solutions were prepared for titration essentially as outlined by Schroeder ( 8 ) . Desired amounts of standard 0.0200011' hydrazine sulfate solution were measured into a 100-ml. beaker and diluted t o 25 ml. with distilled water. The solution was adjusted to about p H 7 by making it alkaline with 0.1s sodium hydroxide, using phenolphthalein as the indicator, and then just discharging the pink color with 0.05.Y hydrochloric acid. Twenty-five milliliters of 95% ethyl alcohol were then added, followed by 0.08 gram of tetrahydroxyquinone. The beaker was next placed over a light beam from the light boy and the sulfate was titrated with standard barium chloride solution measured with a microburet, until the transmitted color changed from green to blue. The titration was made in subdued light, preferably in a semidark room to facilitate observation of the color change
TITRATION CURVES OF LOW SULFATE CONCENTRATIONS
500
400 WAVE
LENQTH
-
600
MP
I
OC 0 2 4 rnl.0.02000 N H Y D R A Z I N E S U L F A T E
Figure 2. Combined transmittance curves for Wratten filter No. 45 and tetrahydroxyquinone
Figure 3. Calibration curves for barium chloride solutions
1. Filter only 2. Before end point 3. After end point
1. 0,004*Vbarium chloride 2. 0.021V barium chloride 3. 0.1N barium chloride
These photometric titration curvecorroborate and help explain the anomalies found in the visual titration for the lower range of sulfate concentration.
trations of barium chloride solution is shown in Table 11. No significant variation between the different strengths n-as found.
EFFECT OF TITRANT CONCENTRATION
INTERFERENCE OF FORElGN IONS
Nuch of the work on the sulfate titration ( 4 6 , 7 , 2 1 ) has employed either 0.01 or 0.02N barium chloride as the titrant. For the current investigation 0.004. 0.02, and 0.1N barium chloride solutions were used to study the reldtive efficiency of the titrant over a wirde range of sulfate concentrations. The 0.1A7 solution gave sharper end point; and could be used over a greater range of sulfate concentrations than the other concentrations. Even with the lesser amounts of sulfate, the 0.1ssolution was more effective, giving a satisfactory visual end point with 0.3 mg. of sulfate, IThile the 0.02 and 0.004.Y r+ agents Ivere satisfactory only down to 0.4 and 0.6 mg. of sulfate, respectively. The main disadvantage of the stronger reagent is the small volume oi titrant to be measured.
Several of the heavier metal cations give colored precipitates with tetrahydroxyquinone. X a n y osidizing agents react with tetrahydrosyquinone. The excess peroxide from the Parr bomb fusion or catalytic combustion of organic compounds must be destroyed before sulfate is estimated by using this indicator for end point detection. Hallett and Kuipers ( 4 ) noted end point fading with tetrahydroxyquinone even after hydrogen peroxide in solution had apparently been destroyed. A difficulty common to all barium sulfate methods for sulfate determination is the coprecipitation of various cations and anions. Experiments of the authors indicated that the order of interference of several anions was phosphate > fluoride > nitrate > chloride, the first three being particularly significant. These results confirm similar n-ork of Fritz and Yamamura ( 2 ) when using Thorin as the indicator. Recently Fritz and others ( 2 , 3) eliminated both
PRECISION AND ACCURACY
An indication of the precision aiiil accuracy attained with three concen-
L
Table II.
BaC12, N 0 004
Analytical Results Sulfate, M g .
Taken 4 80
Found 4 80 4 85 4 86 4 83 4 79 4 86
4 79 4 80
1Iean Std. dev. Acciirncv, % 4 80
0.02
4 82 0 020 0 4 4 76
4 78 4 80
4.81 4.76 4 79 2 76 4.76 Mean Std. dev Accuracy, cTc 19.20
0.1
4 78
0 020 0 4
19.15 19.10 19.10 19.20 19.10 19 20
Mean 19.14 Std. dev. 0.049 Accuracy, c“ 0 3
-4
6 W V
5
cationic and anionic interference through the use of ion exchange columns to separate sulfate from the interfering ions. This resulted in increased accuracy in the determination of sulfate.
-
0.6 -
m
a
-
-
0 v)
m
a
-
0.21 ml.0.l
I
I
I
LITERATURE CITED
I _
I
(1) Eastman Kodak Co., Rmhester, h-. Y.. “Wratten Filters, 1951. (2) Fritz, J. S., Yamamura, S. S., ANAL. CHEJI.27, 1461-4 (3955). (31 Fritz, J. S., Yamamura, S. S., Richard, 11. J., Zbid., 29, 158-60
N BARIUM C H L O R I D E
Figure 4. Photometric titration curves for 0.1N barium chloride vs. 0.02000N hydrazine sulfate
119571
(4) Ha\l?eii, L. T., Kuipers, J. R., IND. E X G . CHEhI., ANAL.ED. 12, 360-3
1. 0.00 ml. hydrazine sulfate 4. 0.40 ml. hydrazine sulfate 2. 0.10 ml. hydrazine sulfate 5 . 0.60 ml. hydrazine sulfate 3 . 0.30 ml. hydrazine sulfate 6. 1.00 ml. hydrazine sulfate
(1940). (5) Lee, S. W., Wallace, J. H., Hand, \Ir. C.. Hannav. N . B.. Zbid.. 14. 839-40 (1942): (6) Mahoney,‘ F.. Mitchell, J. H., Mahoney, J . P.. Zbid., 14, 97-8 (1946). ( 7 ) Ogg, C. L., Willits, C. O., Cooper, F. J.. - 4 s . 1 ~ CHEJI. . 20, 83-5 (1948). (8) Schroeder, \IC., -.IND.ENG.C m a r . , ANAL. ED. 5 . 403-6 (1933). ( 9 ) Sheen, R. T., Knhler, H. L., Zbid., 8, 127-30 (1936). (10) Storlazzi, 11. S., Ransom, V. R., Environments1 Research Laboratorv, Cniversity of Washington, private communication (11) Sundberg, 0 . E., Roger, G. L., ISD. EYG.CHEV... ~ S I L . ED. 18, 71923 (1946). (12) Kalter. R. S . . 4s.11.. CHEM. 22. 1332-34 (1050). ~~~
Table I. 0 02000.~
Hydrazine Sulfate, 311. 0 00 0 10 0 20 0 0 0 0
1 1 2 3 4
30
40 50
60
00 50 00 00 00 5 00
Calibration Data for Barium Chloride Solutions
Equivalrnr Sulfate. Mg. 0 0 0 0 0 0 0 0
00 10
19 29 38 48 58 96
1 44
1 92 2 88
3 84 4 80
0 004.10 40 0 55 0 50 0 30 1 10 1 50 3 60 5 60 8 25 10 65 15 85
20 70
25 95
Barium Chloride, M1. 0 02s 0 12 0 14 0 0 0 0 0
1 1
2
3
13
13
53
62 71 12 65 16 18
4 18 5 22
0 0 0 0 0 0 0 0 0 0 0 0 0
15 03 02 02 09 11
13 15 24 34 45
66 85 1 05
~~
RECEIVED for review Julj- 3, 1956. Accepted April 23, 1957. VOL. 29, NO. 9, SEPTEMBER 1957
1355
