751
V O L U M E 2 3 , NO. 5, M A Y 1 9 5 1 a t least closely approximates this condition. Table X shows that for several samples of pteroylglutamic acid purified under different conditions, analyses for water (determined by the Karl Fischer method), paminobenzoylglutamic acid, and pteroylglutamic acid have been obtained which have added up to practically 1 0 0 ~ o .Although it is fully realized that this is not conclusive evidence, i t is strongly indicative that the present method has no systematic errors. In addition, the values for the purified samples agree with those obtained by a method involving consumption of nitrous acid. This has been reported t o form a nitrosamine in the 10 position ( 4 ) . (The details of this method will be the subject of a later paper.) Both the colorimetric and the titrimetric methods depend upon the measurement of the secondary amine in the 10 position. ACKYOW LEDGNl ENT
The authors wish to thank L. A. Komwowski, S . K. Fiess, and G. 4. Allen for carrying out portions of the experiment’al work reported in this paper. LITERATURE CITED
( 1 ) Alfrey, V., Teply, L. J., Geffen, C., and King, C. G., J . BioZ.
Chem., 178, 465-81 (1949).
(2) Bratton, A. C., and Marshall, E. K., Jr., Ibid., 128, 537-50 (1939). (3) Carpenter, K. J., and Kodicek, E., Biochem. J . , 43, ii (1948). (4) Cosulich, D. B., and Smith, J. M.,Jr., J . Am. Chem. Soc., 71, 3574 (1949). (5) Dennis, L. M., and Shelton, R. S., Ihid., 52, 3128 (1930). (6) Duncan, J. B., and Christian, J. E., J . 4 m . Phawn. rlssoc., Sci. Ed., 37, 507-9 (1948). (7) Glaeko, -1.J., and Wolf, L. hf., Arch. Biochem., 21,241-2 (1949). (8) Hutchings, B. L., Stokstad, E. L. R., Boothe, J. H., Momat, J. H., Waller, C. W., dngier, R. B., Semb, J., and Subbarow, P., J . Bid. Chem., 168,705-10 (1947). (9) “International Critical Tables of Sumerical Data, Physics, Chemistry and Technology,” Vol. 2, p. 589, Kew York, McGraw-Hill Book Co., 1929. (10) Kienle, R. H., and Stearns, E. I., Instruments, 20, 1057-63 (1947). (11) Mader, W.J., and Frediani, H. A., ANAL. CHEM., 20, 1199-1201 (1948). (12) Mellon, B.1. G., “Analytical Absorption Spectroscopy,” New York, John U’iley & Sons, 1950. (13) ”Pharmacopeia of the United States of America,” 14th rev., pp. 250-1, Easton, Pa., Mack Publishing Co., 1950. (14) l-illela, G. G., Hospital, 0 (Rio d e Janeiro), 30, 755-8 (1946). RECEIVED November 15, 1950. The spectrophotometric terminology employed in this paper is t h a t recommended in “Analytical hbsorption Sprctroscopy” (f2).
Determination of Magnesium in Plant Tissue with Thiazole Yellow H. Y. YOUNG AND R. F. GILL Pineapple Research Institute and Hawaiian Pineapple Co., Honolulu, Hawaii The object of this work was to find a precise and accurate direct method for the determination of magnesium in plant tissue which w-ould eliminate time-consuming preliminary separations of cations. A reliable procedure was obtained by utilizing the color reaction of magnesium with thiazole yellow under alkaline conditions. The color was stabilized with polyvinyl alcohol and interferences were eliminated or compensated for by the addition of hydroxylamine, copper, aluminum, calcium, manganese, and phosphate. Results show increased sensi-
T
HE direct colorimetric determination of magnesium with thiazole or titan yellow in plant or soil extracts has been modified in many ways in recent years ( 2 , 6-8). Essentially the modifications consist of the use of a stabilizing agent t o prevent precipitation of the magnesium-dye lake, the addition of a reducing 01 complexing agent such as hydroxylamine, and the addition of various interfering ions a t a concentration above which further modelate increments would not result in changes in color intensity. In the authors’ experience with stabilizing agents, starch and gum ghatti solutions have not proved satisfactory largely because colored solutions prepared with them lacked sensitivity and reproducibility. The use of polyvinyl alcohol as recommended by Ellis (3) 11-as tried and the results confirmed the experience of Heagy (4)who found it greatly increased sensitivity on serum samples. Further gain in sensitivity was attained by increasing the alkali concentration; under this condition polyvinyl alcohol was stable and showed no off colors in contrast to starch and gum ghatti which darken to an orange-amber shade in concentrated alkali. The use of a high concentration of alkali has the additional advantage that variation in acidity of the ash solutions results in relatively slight variation in alkalinity of the final solutions for colorimetry.
tivity and reproducibility overothersimilarrnethods. Standard titrimetric or gravimetric procedures for the determination of magnesium are time-consuming and expensive. A dependable direct method using small amounts of sample containing semimicro amounts of magnesium has obvious advantages. Inasmuch as direct flame photometric procedures are now available for the determination of potassium and calcium, a direct method for magnesium will make possible the rapid estimation of the three major cation components in plant tissue.
The depressing effect of copper as noted by Willson and Wander (9)proved to be a necessary factor for increased sensitivity in the proposed method. Although the color intensities of blank and sample were less in the presence of 15 micrograms of copper, absorbance (optical density) ( 5 )of the sample less that of the blank was tn-o to three times greater than that obtained without copper. The method described herein, therefore, proposes the use of polyvinyl alcohol as the stabilizing agent, an increased alkali concentration, and the incorporation of 15 micrograms of copper into the compensating solution prescribed by Peech and English ( 7 )a3 modified by Drosdoff and Kearpass ( 2 ) METHOD
Apparatus. Either the Evelyn photoelectric colorimeter with filter No. 515 or 540 or the Klett-Summerson photoelectric colorimeter with filter No. 54 has been employed satisfactorily for this determination. Carefully matched tubes were used. Reagents. 1. Hydrochloric acid, 3 N 2. Sulfuric acid. concentrated 3. Hydrogen peroxide, 30% 4. Hydroxylamine hydrochloride, 1 70 (weight/volume) 5 . Compensatin solution, grams per liter: Calcium chlorifie (CaCI2), 1.40
ANALYTICAL CHEMISTRY
752 Aluminum sulfate [A12(SO,)3.1SH20], 0.37 Manganous sulfate (MnS04.H20),0.16 Sodium phosphate (Ka3P04.12H2O), 0.70 Copper sulfate (CuSOa.5HzO), 0.059 Hydrochloric acid, concentrated, 5 ml. 6. Polyvinyl alcohol, Du Pont, Elvanol, Grade 71-24, 2 % (weight/volume). Mix 20 grams with about 400 ml. of water, heat to about 90" C., and stir until dissolved. Cool, dilute to 1 liter, and filter if not clear. Keep in refrigerator. 7 . Mixed reagent. Mix equal parts of 4, 5, and 6 just before use. 8. Thiazole yellow (General Aniline Works, Inc., Albany, X. Y.), 0.02% (weight/volume). Prepare fresh weekly and keep in dark bottle. A concentrate of 0.5% in 50% ethyl alcohol in a dark bottle keeps indefinitely. 9. Sodium hydroxide, 10 N . Prepare with low carbonate alkali. 10. Magnesium standard. Dehydrate magnesium sulfatr (MgS04.7HzO)by heating a t 300" C. for 7 hours or until constant weight is obtained. Prepare a standard stock solution containing 1 mg. of magnesium per ml. by dissolving 1.2375 grams of the anhydrous magnesium sulfate in w t e r and diluting to 250 ml. Dilute to prepare a standard containing 10 micrograms of magnesium per ml. Preparation of Sample. DRY ASH. Ash 10 grams of fresh or 1 gram of dry tissue in a 60-ml. porcelain crucible a t 600" C. for 1 hour or until carbon-free. Dissolve with 15 ml. of 3 LV hydrochloric acid, heating if necessary; transfer to a 100-ml. volumetric flask and dilute to volume. Filter, pipet 10 ml. to a 50-ml.~,volumetric flask and dilute to volume. This is designated sample." The rest of the filtrate may be used foi, other determinations. WET ASH. Transfer 2 grams of fresh or 0.2 gram of dry tissue, to a 200-ml. Kjeldahl flask with about 20 ml. of water, add 4 nil. of sulfuric acid, and digest until fuming starts and the mixture is dark. Cool, add 11 drops of 30% hydrogen peroxide, and digest, for a few minutes or until colorless. Cool and dilute to 100 inl. This is designated "sample." Determination. Pipet into a colorimeter tube a 5-ml. aliquot of sample containing not more than 40 micrograms of magnesium when using filter No. 540 with the Evelyn colorirnetcr or S o . 54 with the Klett instrument, or 25 micrograms when using filter No. 515 with the Evelyn instrument. Add 3 ml. of mixed reagent and mix by swirling (Evelyn tube) or stirring with a thin rod having a horizontally flattened end (Klett tube). Introduce 1ml. of thiazole yellow reagent, mix well, and then add 2 ml. of 10 N sodium hydroxide. Mix thoroughly. After adding thiazole yellow, the sample should not be allowed to stand; the thiazole yellow should be followed immediately with the sodium hydroxide. Allow to stand 10 minutes, tap the tube gently if air bubbles adhere to the side of the tube, and read absorbance in a photoelectric colorimeter using a green filter. Set the instrument a t zero absorbance with a blank prepared with an amount of acid approximately equivalent to that of the sample. Samples containing relatively high salt content, a condition seldom encountered, may develop slight turbidity after 10 minutes. Jn this case read the absorbance after 5 minutes, being careful to set the t)lank accurately a t the same time interval. Inasmuch as the addition of reagents requires less than 1 niinUte, a group of ten samples may be prepared in 10 minutes and transmission readings taken in another 10 minutes. Thus the method is ext'remely rapid. Transfer aliquots of the 10 micrograms per ml. of standard to colorimeter tubes, dilute to 5 ml., and proceed with the det'ermination. Prepare a standard curve by plotting the data obt;iinml. A straight-line relationship exists between absorbance anti concentrat,ion of magnesium.
Table I.
Effect of Polyvinyl Alcohol Concentration on Color Stabilization
(Solutions contain 5.0 ml. of compensating solution, 1 nil. of :% hydroxylamine hydrochloride, 1 ml. of 0.1% thiazole yellow. and 5 ml. of 2.5 N sodium hydroxide in addition t o polyvinyl alcohol and nisgnesium) Polyvinyl Alcohol. Mg, r(5O M I . hll./50 MI. of Mixture of 1Ilxture Result 125 Red ppt. 0.25,1% Red gpt. 125 0 . 5 , 1% 125 Red ppt. 1 . 0 , 1% 125 Turbid 3 . 0 . 1% 125 5 . 0 , 1% Clear 5 . 0 , 2%" 250 Clear 250 Clear 5 . 0 , 4% Eq>iivalentt o 1 ml. of 2% per 10 ml.
Table 11. Effect of .ilkali Concentration on Color Intensity (15 micrograms of AI& 10 ml. of mixture, Evelyn phocoelectnc colorimeter, filter No. 515)
S a O H , me. Absorbance
Effect of Polyvinyl Alcohol Concentration on Color Stabilization. Polyvinyl alcohol concentration of 1 ml. of 0.1% solution in a volume of 12.5 ml. as prescribed by Heagy ( 4 )for plasma or serum extracts, was inadequate in preventing precipitation of the magnesium-thiazole yellow lake with plant ash solutions, owing probably to the comparatively higher salt content, particularly after addition of the compensating solution. Increasing polyvinyl alcohol concentration gave the results shown in Table I. One milliliter of 2% polyvinyl alcohol was taken as optimum for the prescribed procedure where a total volume of 11 ml. was finally obtained. The clarity of the solutions appeared to be further improved later when an increased sodium hydroxide concentration was used.
12 0 396
16
0 4:;'
20 0 4ti7
Sodium Hydroxide Concentration. Inasmuch as variable acidity of samples appeared to cause erratic results, the effect of alkali concentration was ascertained as shown in Table 11. In the presence of polyvinyl alcohol, the magnesium-thiazole yellow color apparently increases markedly in intensity with increase in alkali content. Twenty milliequivalents of sodium hydroxide were taken as the optimum quantity for an 11-ml. volume; a t this concentration slight variation in acidity of the sample causes negligible change in color intensity. This compares with 2.5 to 3.0 me. generally used. The high alkalinity proposed is apparently suitable only when polyvinyl alcohol is employed as the protective colloid. With starch or gum ghatti the color formed is of an orange-amber shade rather than red.
400
2 300
n 6
W
p
0
W
L W
5 250
8 8200 1
100 RESULTS AKD DISCUSS103
6 0 364
3
0 213
1
3
CU ADDED, MICROGRAMS
Figure 1.
Effect of Added Copper on Color Intensity Filter No. 54. 11-ml
Effect of Copper. The depressing effect ot' copper on the intensity of the magnesium-thiazole yellow color as noted by Willson and Wander (9) is confirmpd. However, under the conditions of the proposed method, the presence of copper is a unique advantage because apparently the color depression is in inverse proportion to the magnesium present. The net effect in the presence of 15 micrograms of copper is a depression of approximately 66% of the blank color and a depression of only 35% with 25 mi-
V O L U M E 2 3 , NO. 5, M A Y 1 9 5 1
753
vrograms of magnesium as shown in Table 111. At least a twofold increase in sensitivity is obtained. The effect of adding cyanide as suggested by Killson and Kander is the complete elimination of the effect of copper and iradings become high and sensitivity low as in the case where copper is absent (Table 111). The addition of cyanide, therefore, i-. not prescribed.
DPROPOSED METHOD
MG,MICROGRAMS/II ML.
Figure 2. Comparison of Proposed Method and Drosdorf and Nearpass Method Evelyn photoelectric colorimeter, filter So. 515
The increased sensitivity afforded by copper and polyvinyl alcohol with high alkali concentration is not additive but is dependcnt on both factors-that is, in the presence of copper, substituting starch together with low alkali concentration for polyvinyl alcohol with high alkali concentration does not increase sensitivity. Vice versa, without copper, polyvinyl alcohol n-ith high alkali concentration also does not increase sensitivity. Figure 1 shorvs the effect of adding copper to magnesium standards which contain approximately 3 mici ograms of copper. S e t absorbance reaches practically a constant level a t 10 to 12 micrograms of added copper (total 13 to 15 micrograms). Above this point, an error within -3.1% Kill occur if the ratio of copper t o ni:rgiwiuni is not greater than 1 to 2.5 as shown iii Table IV
Table 111. IIg, y
Cu,
y
Effect of Copper on Absorhance and Sensitivity Reading'
Net Reading
Sensitivity Gainb
DFprmslonC
qepress o n , %d
.4 correction based on Table 117may be applied to results obtained on relatively high copper samples. This has been unnecessary on pineapple tissue. The proposed method eliminates interference from small amounts of copper when present in distilled water stored in partially corroded, tinned-copper tubing and tanks. Inasmuch as the same water is employed for standards and samples, any color-depression will be uniform throughout. Comparison of Colorimetric Methods. The increased sensitivity of the proposed method is further shown in Figure 2 which represents a comparison of the method of Drosdoff and Searpass and the proposed procedure on standard solutions of magnesium. The Evelyn green filter S o . 515 gives absorbance readings higher than S o . 540 by the proposed method, whereaj the difference is slight by the method of Drosdoff and S e a r p a s . Even with the Evelyn filter S o . 510, absorbance by the proposed method is more than tv-ice that of Drosdoff and Searpass. With samples having a wide raiige of magnesium conceiiti:ition, filter No. 540 is more practical than S o . 515. Effect of Manganese. Manganese develops a color nith thiazole yellow similar to that of magnesium. The addition of hydroxylamine and the incorporation of 50 micrograms of nianganese in the compensating solution as suggested \I>- Drosdoff and Xearpass keep manganese interference to 3 niinimuni. However, an error within +5.570 will occur if the ratio of manganese to magnesium is not greater than 1 to 2.5 a3 shown in Table V. With samples containing a higher proportion of manganese a correction based on Table 1-may be applied to the magnesium values. Table V.
Effect of 3Ianganese above 50 Micrograms Mn
hlg, Y
25 25
Added, 0
Total RIn, y 50
y
10
60
Net Readingsa 180 190
Increaae
%, ...
5.5
Ratio Added l l n t o Rlg
... 1to?,5
25 20 i0 198 10. 3 1 to 1 . 2 25 30 80 204 13.3 1 to 0.8 a Klett-Summerson instrument, te>t t u b e , filter T o . 5 4 , l l - : i i l . mixture, 24 micrograms of Cu.
1
,
Figure 3.
I
I
I
15 20 25 30 TIME, MINUTES Stability of Magnesium-Thiazole Yellow Color
5
IO
25 micrograms magnesium in 11-ml., Evelyn phntnelectric colorimeter, filter No. 540 "
IilPtt-Summerson instrument, test tube, filter No. 54. Il-ml. mixture. _103 .
Table 11'. LIg, 5
25
cu Added, 0
Effect of Copper above 15 Micrograms Y
Total Cu, Y 15
Nqt Reading" 191 188
Decrease,
%
Ratio S d d e d Cu t o Rlg
20 i:6b 1 t0'S 25 185 3.1 1t o 2 . 5 25 30 179 6.3 1 t o 1.7 ' Reading minus blank containing 15 micrograms copper, Klett-Summerso! instrument, test tube, filter No. 54, 11-ml. mixture. 191 - 188 For example: 1 9 1(100). 25 25
5 10 15
Stability of Color. Figure 3 s h o m the behavior of the magnesium-thiazole yellow color with time. Apparently a t least 10 minutes are required for the color to become stable. Color intensity is practically constant from 10 to 20 minutes. Samples high in salt content may develop slight turbidity after 10 minutes, in which case readings a t the three- to five-minute interval may be taken and compared with standards prepared similarly. The initial high reading is due to the presence of small air bubbles which gradually rise to the surface within the first 3 minutes. Order of Addition of Reagents and Other Effects. The order of addition of hydroxylamine, compensating solution, and stabilizer is not critical. This is followed by the dye and immediately the sodium hydroxide. With certain samples some
ANALYTICAL CHEMISTRY
754 Table VI. 11g Present,
Mg
Sdded, Y
12.5 12.5 12.5 12.5 12.5 12.5
20.0 20.0 18.5 19.6 24.5 11.7
Accuracy
‘
Total 32.5 32.5 31.0 32.1 37.0 15.2
Total Found 33.0 33.0 32.4 32.8 f7.3 -5.6
.4ccuracy,
%
Mean
101,s 101.5 104.5 102.1 100.8 101,6 102
Table 7’11. Comparison of Colorimetric and Titrimetric iMethods ( 1 ) (With 80 samples, mean difference a s % of titrimetric method = + 1 . 7 % ’ standard error = * O , 8%) hlg in Fresh Tissue, % Colorimetric Titrimetric 0.023 0 026 0 Cli 0.015 0.021 0.021 0.025 0 027 n. 027 n 1327 .~ 0.021 0 023 0 023 0.025 0.024 0 025 0,028 0 028 0 032 0.031 0.023 0 022 0.030 0 030 0 029 0.C31 0.030 0 027 0,028 0 025 0.034 0 032 0.031 0 029 0.037 0 036 0.018 0 018 0 018 0 017
loss of color was noted if the sodium hydroxide was not added immediately. The sodium hydroxide should always be added la&. The first three reagents may be combined. This mixed reagent was found in practice to be stable for a month or more but this should be established for the prevailing conditions in the individual laboratory. The dilute dye was also stable for all practical purposes when protected from light and prepared with
copper-free water. Thorough mixing is essential for reproducibility. Contact with rubber tubing or stoppers must be avoided M higher and irregular values result. Contact with the mixed reagent or sample with reagents gave in some cases up to a third greater absorbance. Although the prescribed compensating solution generally overcomes the effect of interfering ions, a further refinement in technique which contributes to greater accuracy involves incorporating into the standards average quantities of ions which are known from previous analyses of similar material to be present in the samples being analyzed. Thus the effect, particularly that of copper or manganese present in samples, may be equalized by adding appropriate quantities of these elements to the standards. Accuracy. The accuracy of the method is shown in Table VI. Magnesium found averaged 102% of that present. Comparison with Standard Method. Table VI1 presents typical comparative results on pineapple leaf tissue with the standard titrimetric procedure in which magnesium is precipitated as magnesium ammonium phosphate and its acid equivalent determined (1). B mean difference of 1.7% with a standard error of ~ 0 . 8 %was obtained on 50 separate samples. LITERATURE CITED
Official .ID. Chemists, “Official and Tentative Methods of Analysis,” 6th ed., p. 36, 1945. (2) Drosdoff, M., and Nearpass, D. C., ANAL.CHEY.,20, 673 (1948). (3) Ellis, G . H., private communication. (4) Heagy, F. C., Can. J. Research, 26E, 295 (1948). (5) Hiskey, C. F.,B s a ~CHEY., . 21, 1440 (1949). (6) Mikkelsen. D. S., and Toth, S. J., J . Am. SOC.Agron., 39,165-6 (1) Assoc.
(1947).
Peech, AI., and English, L., Soil Sci., 57, 167 (1944). Sterges, A . J., and hIacIntire, W. H., AXIL. CHEM.,22, 351 (1950). (9) Willson, A . E., and Wander, I. IT., Ibid., 22, 195 (1950). (7) (8)
RECEIVED November 25, 1950. Published with the approval of the director as Technical Paper 197 of the Pineapple Research Institute of Hawaii.
Rapid 8-Quinolinol Procedures for Determination of Magnesium A. E. WILLSON’ Citrus Experiment Station, Lake Alfred, Fla.
R
EDMOND and Bright ( 5 ) found that calcium precipitated as the oxalate does not retain any 8-quinolinol (8-hydroxyquinoline, oxine) when magnesium is later precipitated with % qainolinol in the same solution. They used an ammonia solution for the precipitation of magnesium with 8-quinolinol and had to make an RIOa separation first because iron and aluminumalsoform inaoluble oxinates under these conditions. Berg ( 1 ) used an alkaline tartrate solution for the precipitation of magnesium, which has been reported ( 2 ) to give a good separation from aluminum b u t n o t from iron. Other workers ( 4 ) have reported that the iron oxinates are very soluble in sodium hydroxide-tartrate solutions. It was apparent from this literature that it might be possible to determine magnesium directly with 8-quinolinol in an alkaline tartrate solution containing sodium oxalate. The conditions for such a procedure have been studied and methods for determining magnesium directly in leaf tissue solutions have been developed. 1
Present ad&-,
Minute Maid Gorp., Plymouth, Fla.
REAGENTS
Reagents for Precipitating Magnesium. Oxalic-Tartaric Acid Solution. Dissolve 3 grams of oxalic acid and 25 grams of tartaric acid in 75 ml. of water. Adjust the final volume to 100 ml. Sodium Hydroxide, 0.25 N . Dissolve 10.0 grams of sodium hydroxide in 1000 ml. of water. Strong Sodium Hydroxide Solution, 10 S. Dissolve 40 grams of sodium hydroxide in 100 ml. of water. Concentrated Hydrochloric Acid. Potassium Cyanide Solution, 5%. Dissolve 5 grams of potassium cyanide in 100 ml. of water. Bromothymol Blue, 0.05% in a 50% ethyl alcohol-water mixture. Alcoholic %Quinolinol, 1%. Prepare a fresh 1% solution of tb quinolinol in 95% ethyl alcohol as required. Calcium Solution. Place 2.5 grams of calcium carbonate in a 1000-ml. volumetric flask. Add about 100 ml. of water and 5 ml. of concentrated hydrochloric acid. When the calcium carbonate has dissolved, dilute the solution to 1000 ml. Reagents for Determining Magnesium in Precipitate. Wash Solution. Mix 500 ml. of 10% ammonium hydroxide with 500 ml. of 95% ethyl alcohol.
