10-minute levels relative to the moisture content of the sample. EXPERIMENTAL METHOD A N D RESULTS
Figure 3 s h o w the reaction train used for the preliminary experiments and for the thermistor bridge-calcium hydride method. The Schwartz tube with the phosphorus pentoxide was replaced with a thermistor cell. The cell and measuring circuit were like those described by Harris and Nash ( 2 ) . Commercial tank nitrogen, dried with Drierite, was used throughout. A uniform flow rate was used for all experiments; this should be adjusted so that it is rapid enough for quick removal of moisture, but not so rapid as to disrupt the calcium hydride charge in the reaction chamber.
The bridge circuit is balanced initially by passing the dry nitrogen gas through the thermistor vessel a t ambient temperature. The resistance of the decade box is recorded as the initial reading when the galvanometer is at zero. For determination of moisture in ammonium nitrate, the nitrogen is passed first through the test sample and then through the thermistor vessel. The maximum deviation in resistance from the initial decade box setting is recorded. For the work reported, this maximum deviation is related to the moisture content of the ammonium nitrate. Figure 4 shows the calibration curve obtained with the thermistor bridge data for the determination of moisture in ammonium nitrate. The moisture values for the test samples were de-
termined by the Karl Fischer titration. All points are within 0.05 unit, as denoted by the dashed line. The sensitivity of the method is good; for a 0.1% change in moisture content there is about a 110-ohm change in resistance. The time for a n analysis is from 6 to 10 minutes, depending on when the maximum change in resistance takes place. The apparatus is easily assembled and cost of equipment is nominal. LITERATURE CITED
( I ) Engelbrecht, R. M., Drexler, Sam, McCov. F. A.. J . Aar. Food ChenL. 4,
786"(1956). '
(2) Harris, F. E., Nash, L. K., ANAL. CHEM.23, 736 (1951).
RECEIVEDfor review October 8, 1956. Accepted January 14,1957.
Spectrophotometric Determination of Boron in High-Temperature Alloys by Quinalizarin Method A. H. JONES Research Staff, General Motors Corp., Detroit, Mich. Boron in high-temperature aHoys can b e determined spectrophotometrically in the range from 0.01 to 0.1 0% by using the characteristic color of the boron-quinalizarin complex in strong No separations are sulfuric acid. necessary, as extraneous absorption due to colored ions is compensated. Titanium interference, which was revealed b y this investigation, might prove serious a t very low levels of boron concentration, but can b e corrected if the titanium content of the alloy is known. Only concentrated sulfuric acid is used to obtain proper acid concentration. Titration and subsequent adjustment of the concentration are not necessary.
B
has been determined in aluminum-silicon alloys (4)and soils and plants (1) by visual comparison of the color formed when quinalizarin and boron react in strong sulfuric acid. Photometric measurement has been applied to the estimation of boron in plant tissue ( 2 ) and corrosion-resistant steels ( 5 ) , also with quinalizarin as the reagent. I n the last instance, iron was removed from a solution of the sample by means of sodium hydroxide precipitation. I n this investigation an attempt was made t o adapt the reaction to the determination of boron in high-temperature alloys in the range of 0.01 to 0.10% by using the objective spectrophotometric approach. This might permit ORON
accurate measurement of the colored complex in the presence of highly colored ions such as those formed by chromium, nickel, and cobalt, and thus eliminate the need for their removal. Smith ( 4 ) states that none of the common metals spoil the test, and Berger and Truog (1) state that the only known substance that is similar to boric acid in causing a color change with quinalizarin is germanic acid. The method of Berger and Truog was applied t o low alloy steels with no separations (3). Complete solution of the sample is accomplished by fusing fine chips with sodium peroxide in an iron crucible and acidifying a water leach of the melt with sulfuric acid. EXPERIMENTAL
Special Apparatus. Beakers, storage bottles, Erlenmeyer flasks, and condensers made of boron-free glass are available commercially, and are recommended for use wherever possible. The Erlenmeyer flasks and reflux condensers should be equipped mith ground-glass joints. Glass-stoppered, 2-ounce bottles made of soft glass are used for color development. The Beckman Model DU spectrophotometer with 5-cm. cells was used for this investigation. Reagents. Sodium peroxide, reagent grade. Sulfuric acid, c.P., spwific gravity, 1.84. (Some brands of sulfuric acid give a very high blank and should not be used. Those from Du Pont
and Bakm BE Adamson are satisfactory.) Sodium sulfite, c.P., anhydrous. Quinalizarin stock solution. Dissolve 120.0 mg. of quinalizarin in 1 liter of sulfuric acid. This solution is stable indefinitely if kept free from water. Quinalizarin reagent solution. Dilute 25.00 ml. of the stock solution to 500 ml. nith sulfuric acid and mix. Standard boron solution. Dissolve 285.8 mg. of boric acid in water and dilute to 500 ml.; contains 100.0 y of boron per ml. Standard titanium solution. Transfer 1.657 grams of titanium dioxide (KBS standard sample 154a, dried a t 105' C. for 2 hours) to a 500-ml. Erlenmeyer flask, add 10 grams of ammonium sulfate and 22.0 ml. of sulfuric acid, and heat until all the titanium dioxide is dissolved. Cool and rapidly pour the solution into 450 ml. of cool water which is vigorously stirred. Rinse the flask with sulfuric acid (5 to 95). Transfer the solution to a 500-ml. volumetric flask, cool to room temperature, dilute to volume, and mix. This solution contains 2.00 mg. of titanium per ml. Calibration Solutions. Transfer 140 ml. of water t o each of five 200-ml. volumetric flasks. Immerse the flasks in cold water t o prevent breakage due t o thermal shock, add 40.0 ml. of sulfuric acid with swirling and cool to room temperature. For evaluation of the titanium interference, pipet 50.00 ml. of standard titanium solution into each of five 200-ml. volumetric flasks. Add 90 ml. of water and mix by sm-irling. VOL. 29, NO. 7, JULY 1957
1101
Immerse the flasks in cold water, add 38.0 ml. of sulfuric acid with swirling, and cool t o room temperature. A 2.00nil. aliquot of each of these solutions contains 1.00 mg. of titanium. Dilute one solution in each set to volume with water and mix. To the others, add 2.00, 4.00. 6.00, and 8.00 nil. of standard boron solution by means of a semimicro buret. Dilute to volume with water and mix. Color Development. Pipet 2.00 ml. of each of t h e calibration solutions into a dry, glass-stoppered, 2-ounce bottle. Add 25.00 ml. of quinalizarin reagent solution from a pipet, swirling t o prevent breakage due t o thermal shock. Stopper the bottle a n d mix by swirling. Allow t h e solution t o stand for 2 hours i 5 minutes. Measure t h e absorbance or transmittance a t 615 mp in 5-cm. cells, using a reference solution prepared by pipetting 25.00 ml. of sulfuric acid into a 2-1111. aliquot of the calibration solution containing no titanium or boron. Calibration Curves. Calculate t h e absorbance (log 1/T) from t h e transmittance; then plot t h e difference between t h e absorbance for each level of boron concentration a n d t h e absorbance of the solution containing no boron or titanium against boron concentration on t h e same rectangular coordinate paper. Two curves, A (no titanium) and B (1 mg. of titanium present) will result (Figure 1). For iise in obtaining the correct concentration of boron when titanium is present, a nomograph can be constructed as follows: Use a sheet of paper large enough for accurate reading of the nomograph; 16 inches by 22 inches is suggested. Draw a vertical line near the left-hand edge of the paper. Divide the line into eight equal parts and designate the divisions as micrograins of boron, 0 to 8. Mark off each division in tenths, and each of those in fifths. On the calibration graph (Figure l), determine the point (micrograms of boron) where the two curves cross. From this point on the nomograph scale for micrograms of boron, draw a horizontal line perpendicular to the scale. Connect the horizontal line with a vertical line on the right side of the paper. Bisect the horizontal line and designate the point of bisection as 0 mg. of titanium. K i t h a straightedge, connect each point on the left-hand scale representing 0.1 y of boron through the zero point on the horizontal line, with the right-hand line. Designate the point of intersection of the straightedge with the right-hand lint with the appropriate absorbance value. as read from calibration curve A . Use these 80 values on the absorbance scale to interpolate values for even divisions on the scale, many of vhich will correspond exactly with those already obtained. From curve B on the calibration graph, obtain the absorbance values for each microgram of boron. With a straightedge, connect each of these values of the absorbance scale of the nomograph lyith its corresponding value on the micrograms of boron scale and 1 102
ANALYTICAL CHEMISTRY
bl rocra-,
01
EX*,
Figure 1 . Calibration curves showing interference of titanium A. B.
No titanium 1 mg. of titanium
mark the intersection of the straightedge with the horizontal line. Because of method error, the points of intersection will not coincide exactly; therefore, locate the point on the line which represents the average and erase the other points. Designate this point as 1.0 nig. of titanium. Divide the distance betlveen this point and the point designated as 0 mg. of titanium into tenths. Procedure. Transfer a suitable weight (calculated t o contain not moFe t h a n 0.3 mg. of boron or 50 mg. of titanium) of the finely divided sample t o a 25-ml. iron crucible containing 8.0 grams of sodium peroxide. Mix t h e sample and peroxide intimately by stirring with a small spatula or iron rod. Cover t h e crucible and heat moderately until the peroxide is molten. When the vigorous action ceases, remove the cover and swirl the melt for 5 minutes, increasing the heat until the melt has an orangered glow. Cool the melt, place the crucible and the crucible cover in a 250-ml. beaker, and corer with another 250-ml. beaker. Add 25 ml. of water and, n-hen the vigorous action subsides, digest until disintegration of the melt is complete. Remove and rinse the beaker cover, then remove the crucible and crucible cover and rinse them n.ith a minimum of mater, collecting the mshings in the beaker Transfer the solution and insoluble matter to a 300-ml. Erlenmeyer flask. connect to a reflux water condenser, and immerse the flask in cold water. Measure 26.0 ml. of sulfuric acid into a dry 50-nil. graduate. Add the acid to the sample solution in m a l l increments (approximately 5 ml.) by pouring it through the condenser. Reserve the graduate. Remove the cold water bath and heat the flask until the solution has boiled gently for about 5 minutes Cool the flask to room temperature. To transfer all of the sulfuric acid to the flask, rinse the graduate and the condenser with not more than 15 nil. of water. S\virL,the contents of the flask and dis-
connect the condenser.aTo the flask add enough sodium sulfite (solid) to reduce all the dichromate ions, plus a few milligrams in excess. Allow SUEcient time for complete reduction of the dichromate. Transfer the contents of the flask to a 100-ml. volumetric flask, dilute to volume with water and mix. Prepare a blank by performing all of the steps of the procedure described above, with the sample omitted. When sodium peroxide is fused in the absencix of sample, the iron crucible is much more readily attacked. An attempt should be made to obtain only as much iron in the fusion as would be obtained when fusing the sample. Pipet 2.00 ml. of the sample solution to each of two glass-stoppered, 2-ounce bottles; t o one of them add 25.00 ml. of the quinalizarin reagent solution with swirling. Stopper t h e bottle and mix by swirling. To the other aliquot, to be used as the reference, add 26.00 ml. of sulfuric acid. Allow the solutions to stand for 2 hours + 5 minutes, then measure the absorbance or transmittance a t 615 mp in 5-cm, cells. Follow the procedure described above to develop the color in the blank solution. Calculations. Convert each transmittance value t o absorbance and subtract t h e absorbance value of t h e blank solution from t h a t of t h e test solution. If no titanium is present in t h e sample, read the corresponding weight of boron from calibration curve A (Figure I). If titanium is present, calculate the milligrams of titanium i n t h e 2.00-ml. aliquot taken for color development. On t h e calibration graph, locate a point opposite the net absorbance value and between curves A and B. Determine the horizontal locus of the point as follows: hleasure the distance between the two curves and multiply this distance by the number of milligrams of titanium in the 2.00-ml. aliquot. Plot the point a horizontal distance from curve A toward curve B equal to the result. Read the corresponding weight of boron in micrograms If a nomograph is used, micrograms
Immerse the flasks in cold water, add 38.0 ml. of sulfuric acid with swirling, and cool to room temperature. A 2.00nil. aliquot of each of these solutions contains 1.00 mg. of titanium. Dilute one solution in each set to volume with water and m i x To the others, add 2.00, 4.00, 6.00, and 8.00 nil. of standard boron solution by means of a semimicro buret. Dilute to volume n i t h mater and mix. Color Development. Pipet 2.00 ml. of each of t h e calibration solutions into a dry, glass-stoppered, 2-ounce bottle. Add 25.00 ml. of quinalizarin reagent solution from a pipet, swirling t o prevent breakage due t o thermal shock. Stopper t h e bottle a n d mix by swirling. Allow t h e solution t o stand for 2 hours 1 5 minutes. hIeasure t h e absorbance or transmittance a t 615 mp in 5-cm. cells, using a reference solution prepared by pipetting 25.00 ml. of sulfuric acid into a 2-1111. aliquot of the calibration solution containing no titanium or boron. Calibration Curves. Calculate t h e absorbance (log 115") from t h e transmittance; then plot t h e difference between t h e absorbance for each lcvel of boron concentration and t h e absorbance of t h e solution containing no boron or titanium against boron concentration on t h e same rectangular coordinate paper. Two curves, A (no titanium) and B (1 mg. of titanium present) will result (Figure 1). For use in obtaining the correct concentration of boron when titanium is present, a nomograph can be constructed as follows: Use a sheet of paper large enough for accurate reading of the nomograph; 16 inches by 22 inches is suggested. Draw a vertical line near the left-hand edge of the paper. Divide the line into eight equal parts and designate the divisions as micrograms of boron, 0 to 8. hiark off each division in tenths, and each of those in fifths. On the calibration graph (Figure l ) , determine the point (micrograms of boron) where the two curves cross. From this point on the nomograph scale for micrograms of boron, draw a horizontal line perpendicular to the scale. Connect the horizontal line with a vertical line on the right side of the paper. Bisect the horizontal line and designate the point of bisection as 0 mg. of titanium. K i t h a straightedge, connect each point 011 the left-hand scale representing 0.1 y of boron through the zero point on the horizontal line, with the right-hand line. Designate the point of intersection of the straightedge with the right-hand line with the appropriate absorbance value, as read from calibration curve A. Use these 80 values on the absorbance scale to interpolate values for even divisions on the scale, many of which will correspond exactly with those already obtained. From curve B on the calibration graph, obtain the absorbance values for each microgram of boron. With a straightedge, connect each of these values of the absorbance scale of the nomograph with its correqponding value on the micrograms of boron scale and 1 102
ANALYTICAL CHEMISTRY
i1,:iqrdri
01
Boron
Figure 1 . Calibration curves showing interference of titanium A. 6.
No titanium 1 mg. of titanium
mark the intersection of the straightedge with the horizontal line. Because of method error, the points of intersection will not coincide exactly; therefore, locate the point on the line which represents the average and erase the other points. Designate this point as 1.0 mg. of titanium. Divide the distance between this point and the point designated as 0 mg. of titanium into tenths. Procedure. Transfer a suitable weight (calculated t o contain not more t h a n 0.3 mg. of boron or 50 mg. of titanium) of the finely divided sample t o a 25-ml. iron crucible containing 8.0 grams of sodium peroxide. Mix t h e sample and peroxide intimately by stirring with a small spatula or iron rod. Cover the crucible and heat moderately until t h e peroxide is molten. When the vigorous action ceases, remove the cover and swirl the melt for 5 minutes, increasing the heat until the melt has an orangered glow. Cool the melt, place the crucible and the crucible cover in a 250-ml. beaker, and cover with another 250-ml. beaker. Add 23 ml. of water and, when the vigorous action subsides, digest until disintegration of the melt is complete. Remove and rinse the beaker cover, then remove the crucible and crucible cover and rinse them with a minimum of water. collecting the washings in the beaker. Transfer the solution and insoluble niatter to a 300-ml. Erlenmeyer flask, connect to a reflux water condenser, and immerse the flask in cold water. hleasure 26.0 ml. of sulfuric acid into a dry 50-nil. graduate. Add the acid to the sample solution in small increments (approximately 5 ml.) by pouring it through the condenser. Reserve the graduate. Remove the cold water bath and heat the flask until the solution has boiled gently for about 5 minutes. Cool the flask to room temperature. To transfer all of the sulfuric acid to the flask, rinse the graduate and the condenser with not more than 15 nil. of water. Bn-irl,the contents of the flask and dis-
connect the condenser.QTo the flask add enough sodium sulfite (solid) to reduce all the dichromate ions, plus a few milligrams in excess. Allow sufficient time for complete reduction of thc dichromate. Transfer the contents of the flask to a 100-ml. volumetric flask, dilute t o volume with water and mix. Prepare a blank by performing all of the steps of the procedure described above, with the sample omitted. When sodium peroxide is fused in the absenct. of sample, the iron crucible is much more readily attacked. -4n attempt should be made to obtain only as much iron in the fusion as would be obtained when fusing the sample. Pipet 2.00 ml. of the sample solution to each of two glass-stoppered, 2-ounce bottles; to one of them add 25.00 ml. of the quinalizarin reagent solution with swirling. Stopper the bottle and mix by swirling. To the other aliquot, to be used as the reference, add 25.00 ml. of sulfuric acid. illlow the solutions to stand for 2 hours *5 minutes, then measure the absorbance or transmittance a t 613 mp in 5-cm. cells. Follox the procedure described above to develop the color in the blank solution. Calculations. Convert each transmittance value t o absorbance and subtract t h e absorbance value of t h e blank solution from t h a t of t h e test solution. If no titanium is present in the sample, read the corresponding weight of boron from calibration curve A [Figure 1). If titanium is mesent, calculate the milligrams o f ' t i t a n i u m in t h e 2.00-ml. aliquot taken for color development. On t h e calibration graph, locate a point opposite the net absorbance value and between curves A and B. Determine the horizontal locus of the point as follows: hleasure the distance between the two curves and multiply this distance by the number of milligrams of titanium in the 2.00-ml. aliquot. Plot the point a horizontal distance from curve A toward curve B equal to the result. Read the corresponding weight of boron in micrograms If a nomograph is used, micrograms
in a concentration of about 92 to 94% sulfuric acid by weight. I n order to achieve this concentration, Berger and Truog’s procedure ( I ) requires fuming sulfuric acid. RIacDougall and Biggs ( 2 ) give several reasons for avoiding the use of fuming sulfuric acid. I n the procedure described here, the desired concentration in the test solution, about 92.6% by weight, has been obtained by the use of C.P. sulfuric acid, which has a minimum concentration of 95.5oJ, sulfuric acid by weight. Figure 3 shows the effect of sulfuric acid upon absorbance. Curve A was obtained using solutions of quinalizarin containing no boron, while solutions containing the same amount of quinalizarin plus 10 y of boron were used to obtain curve B. Curve C lyas obtained by subtracking the absorbance of each solution containing no boron from that containing boron a t each sulfuric acid concentration level and plotting the resultant data. A decrease in the slope of curve C is manifested a t the higher concentration. levels. Above 92%, sulfuric acid concentration can vary without undue variation in the absorbance due to the boron-quinalizarin complex, provided the sulfuric acid concentration of the solution containing boron and of that containing no boron are the same. In order to determine whether the acid concentration of C.P. sulfuric acid is constant enough to provide consistent results, five lots of the acid were selected a t random. Two calibration series were prepared from each lot. The results (Table 111) indicate that variation from this cause is unlikely. Variation in the absorbance of the blank is attributed a t least in part to boron contamination of the sulfuric acid. Error that might be incurred by this variation is eliminated by determining the blank-i.e., the absorbance of the quinalizarin with no each boron or titanium added-for series of determinations. Quinalizarin Concentration. At any level of boron concentration, not only the absorbance of the quinalizarin, but also the net absorbance due to the boron-quinalizarin complex varies with a variation in quinalizarin concentration. This is illustrated in Figure 4, in which are shown calibration curves a t three levels of quinalizarin concentration. Color Stability. This work showed t h a t after the heat caused by the reaction of sulfuric acid with water had been dissipated, the absorbance reached a constant level and remained constant until a t least 2 hours had elapsed after addition of the reagent solution. I n preference to accelerated cooling of the solution, a waiting period of approximately 2 hours is recommended. Smith (4) states t h a t the color, when fully developed, is stable 1104
*
ANALYTICAL CHEMISTRY
Table 111.
Sulfuric
Acid
Lot s o , 1
Typical Calibration Data
Absorbance of
2 3 4 5
Blank
0,618 0.618 0,644 0,646 0.636 0,629 0.631 0.635 0.633 0.628
0.136 0.136 0.138 0.137 0.141 0.138 0.139 0.137 0.138 0.143 0 138 0 005 1 5s
a
Av. absorbance Max. dev. Std. dev., o/c
0,245 0.254 0.249 0.247 0.250 0.247 0.247 0.248 0.250 0.251 0 249 0 005 1 04
0.341 0.344 0.343 0.339 0.343 0.340 0.343 0.342 0.341 0.349 0.343 0.006 0 82
0.414 0.419 0.413 0.412 0.416 0.415 0.411 0.409 0.411 0.427 0.415 0 012 1 30
7m t
Boron Micrcgramr pr B MI.1
Figure 4. Effect of quinalizarin concentration upon absorbance Ail solutions contain 89.8% sulfuric acid 1. 0.095 mg. of quinalizarin 2. 0.1 90 mg. of quinalizarin 3. 0.285 mg. of quinalizarin
almost indefinitely if the solution is kept in a closed container to avoid absorption of water. Interferences. According to Smith (4) fluorides, nitrates, dichromates, and other oxidizing compounds have a bleaching effect upon the color of the boron-quinalizarin complex. I n the proposed procedure these agents either are not present or are reduced by sodium sulfite. Colored ions in the test solution would cause high absorbance values if no compensation were made for them. This compensation is accomplished by using a reference solution which contains an aliquot of sample solution equivalent t o that in the test solution, with acidity conditions the same. Of the elements usually present in high-temperature alloys, only titanium
appears to form a complex similar to that of boron with quinalizarin. However, the effect of titanium upon the absorbance a t 615 mp is small, as illustrated in Figure 1. The fact that the two curves cross in the region of 4 y of boron can be explained by the theory that while the titanium-quinalizarin complex does not absorb greatly a t 615 mp, it does consume quinalizarin. Thus, because less quinalizarin is available to react with boron, the sensitivity is decreased and the slope of the calibration curve is diminished, as shown in Figure 4. This theory is supported by the fact that the absorption peaks of quinalizarin a t 532 and 572 mp (Figure 2) are not so high when titanium is present. Experimentation has shown that calibration curves obtained with 1 mg. of titanium or less always cross a t the same point, around 4 y of boron. Further ex-
periments showed that, if the quinalizarin content were increased, the slopes of thecurves would become steeper and the curves would converge at a higher point. The absorbance values with titanium Iwsent are aa repeatable as those with titmium absent if the quinalizarin content is held constant. This fact is utilized in making correction for titanium interference.
ACKNOWLEDGMENT
The author wishes to acknowledge the cooperation of his coworkers, R. E. Kohn and R. L. Chance, in the preparation of this paper. LITERATURE CITED
(1) Berger, K. C., Truog, E., IND. ENG. CHEM., ‘4NAL.ED. 11,540 (1939).
MacDougall, Daniel, Biggs, D. A., ANAL.CHEX 24, 566 (1952). Rudolph, G. A4.,Flickinger, L. c., Steel 112, 114 (April 5, 1943). Smith, G. S., Analyst 60, 735 (1935 ) . Weinberg, Sidney, Procter, K. L.. Plliller, O., ANAL. CHEM. 17, 419 (1945). RECEIVEDfor review May 9, 1956. Accepted Xarch 27, 1957. Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, Pittsburgh, Pa., March 1957.
Characteristics of Boron-Benzoin Complex Improved Fluorometric Determination of Boron CHARLES E. WHITE and DONALD E. HOFFMAN’ Department o f Chemistry, University o f Maryland, College Park, Md.
b The boron-benzoin complex used in the determination of boron is stabilized a t its maximum intensity for a 20minute period with glycine buffer and for a somewhat longer time with isopropylamine. The complex is decomposed b y extended exposure to ultraviolet light. The maximum fluorescence emission is from 450 to 520 mk and the maximum excitation is from 310 to 400 mp. The complex contains 1 gram atom of boron to 1 mole of benzoin.
B
is an excellent reagent for boron. Details of the qualitative and quantitative procedure for this determination have already been published (6). From a quantitative standpoint a disadvantage in the use of this reagent was the tendency for the fluorescence to reach a maximum intensity in a period of about 3 minutes and then decrease rapidly after 5 minutes. The fluorometer used in the original measurements was one which did not have a shutter imposed between the light and the sample, and permitted the continuous exposure of the sample to the exciting radiation, The fluorescence is stable for a much longer period if exposure of the sample is confined to the brief interval during which a reading is being made. A means of increasing the stability of the fluorescence and other characteristics of the boron-benzoin complex have been determined. ENZOIN
APPARATUS AND REAGENTS
The fluorometer was constructed in Present address, Chemistry Department, University of Delaware, Newark, Del.
0
I
2
3
4
5
6
7
8
9
IO
I1
12
I3
14
1516
TIME -MINUTES AFTER MIXING
Figure 1. Variation of fluorescence intensity with time under continuous irradiation A. 8.
0.1 ml. of 0.6N sodium hydroxide 0.4 ml. of glycine buffer (pH 12.8)
this laboratory and consisted of a sample holder, a 1P28 photomultiplier tube attached to a sensitive microammeter (R.C.A. W.V. S a ) , and a power source for the phototube (4). The usual source of ultraviolet light was a General Electric Co. H100B4 lamp; however, for the excitation spectra the radiation from a xenon arc mas passed through a Bausch & Lomb grating monochromator (KO. 33-8640-01). A Beckman Model DU spectrophotometer equipped with a photomultiplier tube was used to measure the absorption spectra. All inorganic compounds were reagent grade. Commercial 957& ethyl alcohol was redistilled to remove a nonvolatile fluorescent impurity. Boron solution was prepared by dissolving boric acid crystals in water in proper amounts to afford concentrations of 1 to 10 y of boron per ml. and 50 y per ml. Benzoin (Eastman Kodak Co.) n a s recrystallized from benzene and dissolved in redistilled 95% ethyl alcohol to produce qolutions of 0 5 0 2, and 0.137,. Glycine buffer was prepared for a pH of 12.8 by thc method of Clark ( I ) .
Two stock solutions were necessary. Solution A contained 7.505 grams of glycine, 5.85 grams of sodium chloride, and 1 liter of water; solution B contained 2 grams of sodium hydroxide pellets dissolved in 500 ml. of water. For the buffer 50 ml. of A was mixed with 430 ml. of B. These solutions were stored in polyethylene bottles. Isopropylamine (Matheson, Coleman, & Bell Co.) was distilled to remove a nonvolatile fluorescent impurity. A stock solution of quinine sulfate ivas prepared by dissolving 0.01 gram (USP, Jlerck &- Co., Inc., New York) in 1 liter of nater a n d diluting further Rith 0.l.Y wlfuric acid as desired. -4 solution of a fluorescence in the rai’ge of the determinations in progress was used to set the fluorometer for comparabk rtadings over a period of time. Boron-free glassware and quartz test tubes were used in routine work. d l 1 alkaline solutions were stored in polyethylene containers. Glass-stoppered borosilicate glass test tubes, graduated a t 25 mi., n-we used for fluorometric nwasurements. I n thesc cases the solutions wtre never in the test tubes for more than 1 hour because alkaline soluVOL. 2 9 ,
NO. 7, JULY 1957
1105