October 15, 1932
INDUSTRIAL AND ENGINEERING CHEMISTRY
ards Ingot Iron No, 55 with sulfuric acid, as described in the sulfuric acid method above. Determinations 5 to 8 were made by dissolving 10- to 20-gram samples of the iron in sulfuric acid, evaporating, igniting, and volatilizing with hydrochloric acid. Blanks were determined on all reagents used, and any silica found was subtracted. I n no case was a blank of more than 0.2 mg. of silica obtained, whereas in most cases the amount was 0.1 mg. or less. These results show that the true silicon content of this iron is of the order of 0.002 per cent, as compared with 0.0005 per cent or less indicated by dehydration with sulfuric acid. ACKNOWLEDGMENT The writer acknowledges the cooperation of H. A. Bright in the preparation of this paper.
385
LITERATURE CITED (I) Boussingalt, Ann. chirn. phys.,
[4] 22, 457 (1871). (2) Gooch, F. A., and Havens, F. S., Science, [4] 7, 370 (1899). (3) Hillebrand, W. F., and Lundell, G. E. F., “Applied Inorganic Analysis,” p. 549, Wiley, 1929. (4) Lundell, G . E. F., Hoffman, J. I., and Bright, H. A,, “Chemical Analysis of Iron and Steel,” p. 58, Wiley, 1931. (5) Lundell, G. E. F., Hoffman, J. I , and Bright, H. A., I b i d , pp. 258-62. (6) Pinsel, H., Chern.-Ztg., 50, 924-5 (1926). (7) Stadeler, A., Chemikerausschusses der Ver. deut. Eisenhtittenleute, Ber., 52 (1927). (8) Thayer, L. A., IND.ENO.CHEW,Anal. Ed., 2, 277 (1929). (9) Willard, H. H., and Cake, W. E., J . Am. Chern. SOC.,42, 2208 (1920). (10) Wolf, H., and Helingotter, R., Chem.-Ztg., 54, 878-9 (1930).
RECEIVEDApril 8, 1932. Publication approved by the Director, Bureau of Standards
Determination of Boron Spectroscopically J. S. MCHARGUE AND R. K. CALFEE,Kentucky Agricultural Experiment Station, Lexington, Ky. visible. Increase in the boron A new spectroscopic method for the determinac o n t e n t of t h e s o l u t i o n reboron in plant metabotion of boron is described. The procedure consulted in a spectrum of greater lism h a s b e e n a s c e r sists in ashing the plant material at a low temi n t e n s i t y . To measure the tained physiologically by perature in the presence of an excess of a n alkali, relative intensity of the specseveral investigators. Bertrand and dissolving the ash in a solution of citric acid. trum, a number of s o l u t i o n s and A g u l h o n (2, 3, 4 ) h a v e of c o l o r e d s a l t s were tried demonstrated the p r e s e n c e of Boron is separated from the solution of salts by as light filters. A dilute soluthe element in normal tissues converting it into methyl borate and distilling. tion of p o t a s s i u m permangaof v a r i o u s s p e c i e s of plants The methyl borate is burned in oxygen by a nate as well a s s o l u t i o n s of and animals. Jay (7), investispecial type of apparatus and the spectrum several organic dyes were cagating the occurrence of boron absorbed by the addition of a standard solution pable of absorbing the green porin n a t u r a l foods, found it to tion of the spectrum. Potasbe widely distributed. Tests of potassium permanganate from a buret into a sium permanganate was confor boron in plant and animal glass cell containing distilled water and cons i d e r e d the most suitable betissues have, for t h e most verting the permanganate factor to its equivalent cause it can be standardized p a r t , been m e a s u r e m e n t s in boron. Extreme sensitivity was obtained by readily, and is the least subject of t h e d e p t h of color imburning the methyl borate in a n atmosphere of to change under laboratory conparted to standardized strips ditions. of turmeric paper. Details of oxygen. Boron can be determined quantitatively Measurement of the intensity methods u s i n g turmeric, curon samples of plant tissues containing 0.05 to of the s p e c t r u m was effected cumin, and tincture of mimosa 0.3 mg. of boron with a n accuracy of t0.0095mg. by adding 0.01 N p o t a s s i u m blossoms have been described permanganate solution from a by several investigators (4, 6, 6;8). These methods are very sensitive to small quanti- buret to 50 ml. of distilled water in a glass cell placed beties of boron, but are only roughly quantitative. Bertrand tween the spectroscope and the flame, until the band was and Agulhon (9)confirmed their results with turmeric paper just eliminated. Ten milliliters of acid alcohol containing by spectroscopic examination of the color imparted to the a known amount of boron were placed in a 50-ml. Erlenmeyer flask fitted to a short hard glass tube fixed in a glass flame of burning hydrogen. The green color imparted to the flame of burning alcohol chimney (Figure 1). Two glass tubes projecting from the is a very sensitive test for boron. To test for boron in the chimney fixed the distance of the spectroscope from the flame presence of other compounds producing a green flame, Bordas a t 6 cm. and supported a cell of 2 cm. thickness under a buret, and Touplain (6) vaporized the acid solution of the sample The flask was immersed in a constanbtemperature water in methyl alcohol and ignited the vapor at the end of a glass bath, and the temperature was raised until the flame was tube. Boric acid with methyl alcohol in the presence of a 1.5 cm. high. A direct-vision Schmidt and Haensch spectrodehydrating acid forms methyl borate, (CH&BOs. The scope was placed before the cell, and 0.01 N potassium perdifference in the boiling points of the boric acid ester and of manganate solution was added slowly, with stirring, until methyl alcohol is 1’ C., and it has been found that with the first green band of the spectrum was absorbed. It was small amounts of the ester present the boiling point is con- found that a definite normality of potassium permanganate solution 2 cm. thick would just eliminate the lines from the stant. Spectroscopic examination of the flame of the mixed vapors spectrum produced by the presence of a definite quantity of the alcohol and ester showed four bands between the of boron in the solution, other conditions being identical. sodium line and the first blue caesium line, the most inConditions other than the quantity of boron present that tense being close to that of thallium in the green portion of might affect the intensity of the spectrum were investithe spectrum. With less than 0033 gram of boron per gated. It was found (Table I) that the distance of the spectroliter of alcohol, it was found that the bands were not
HE essential nature of
T
386
.Vol. 4, No. 4
ANALYTICAL EDITION TABLEI. MINORFACTORS INFLUENCING SPECTRUM
INTENSITY OF
(Average of 3 exoeriments) POTASSIUM PERMANGANATE
FACTOR
1. Distance of spectroscope from flame 3 em. 6 om." 8 om. 2.
B 0.05 SOLUTION B = d.10 001 NB = 0.5 gram/liter gram/liter gram/liter
ME.
ME.
M1.
4.8 4.6 4.2
7.1 7.0 6.8
10.15 10.1 9.9
7.1 7.0 7.0
10.1 10.1 10.05
7.3 7.0 7.0 6.9
10.7 10.4 10.1 9.95
6.7 7.0 7.2
10.0 10.15 10.3
Position of cell Against chimney 4.6 Against spectroscopeQ 4.6 Intermediate 4.6 3. Height of flame 0.5 om. 5.0 1.0 om. 4.7 1.5 om." 4.65 3.0 cm. 4.5 4. Opening of slit Slight (Na to 3060) 4.1 Moderate (Na to 3053)a 4.6 Wide (Na to 3045) 4.8 a Found to be most satisfactory condition.
FIGURE1. APPARATUS FOR SPECTROSCOPIC DETERMINATION OF BORON A . Spectroscope B . Buret
H. Bath
I . Heater J. Thermometer Mercury bulb Thermostat Regulation Pilot light 0. Condenser
Cell D . Support E. Chimney F . Burner Q. Sample flask C.
K. L. M. N.
.
The practice of adding bDric acid or borates to foods as preservatives has been responsible for the development of several methods for the extraction of boron from plant material. These methods are commonly used in conjunction with volumetric or gravimetric determinations, and were found to be unsatisfactory for the analysis of minute quantities, as small losses occur in the preparation of the sample. It was found that losses may occur on ignition above 800' C. and on evaporation of the acid solution. 12 11 10
d9 scope from the flame affected the intensity of the spectrum, especially with the smaller quantities. This distance was fixed at 6 cm. by glass stops for all determinations. The z position of the cell did not exert any noticeable influence % 6 so long as it was perpendicular to the spectroscope. Dif9 5 ; ferences in the height of the flame resulted in different intensities, the intensity decreasing as the size of the flame increased. The flame must be over 1 cm. in height in order to burn smoothly and quietly. This was accomplished by 2 heating the bath slowly until the top of the flame pias just i opposite the 1.5-cm. mark on the chimney, and adjusting ,. the thermostat for this temperature. The setting of the 0 I 2 3 4 5 6 7 8 9 1 0 slit of the spectroscope also influenced the intensity by varyBoron gms. per l i t e r ing the amount of light admitted. It was opened until the FIGURE2. POTASSIUM PERMANGANATE bands were sharp, and set at the same figure for all determiSOLUTIONREQUIREDTO ELIMINATE nations. Dissolved salts, platinum wire, and other catalysts FIRSTGREEN BANDFROM SPECTRA OF BORON BURNED IN AIR exert a minor influence that is overcome by the presence of 5 per cent phosphoric acid in the alcohol. The effect is due Table I1 shows recoveries from salt mixtures comparable to hastening the formation of methyl borate. The difference in the boiling points of the alcohol and ester to plant ash. did not affect the intensit'y for quantities of boron within OF BORON IN PRESENCE OF Ca, Mg, the range of the method. The amounts of potassium per- TABLE11. EXTRACTION AND K IN HC1 SOLUTION manganate solution required for a given amount of boron TREATMENT B PRESENT B RECOVERED were checked 15 minutes after igniting the vapor. The Gram Gram curve (Figure 2) shows the exact quantities of 0.01 N po- Evaporated (acid), extracted with methyl 0.0016 0.0002 alcohol, and filtered 0.0035 0.0005 tassium permanganate solution necessary to produce suf- Evaporated with &Con, extracted with 0.0016 0.0008 ficient color in 50 ml. of water t o absorb the first green band methvl alcohol. and filtered 0.0035 0.0020 with methyl alcohol containing of the spectrum of known amounts of boron. The position Evaporated 25% NHaOH, extracted with methyl alco0.0016 0.00097 of each point measured on the curve is the average of from hol, and filtered 0.0035 0.00294 Evaporated with methyl alcohol containing three to five determinations. 25% "409, extracted with methyl alcohol 0.0016 0.00167 for 3 hours in Soxhlet tube 0.0035 0.00343 A table showing the boron equivalent of each 0.1 ml. of 0.01 N potassium permanganate solution can be prepared Evaporation of acid solutions on the water bath results by calculations from the data on known quantities of boron for convenience in converting buret readings to grams per in a loss of boron amounting to 75 per cent or more, if very small quantities are present. The addition of an alkali liter of boron.
3;
2'
October 15, 1932
INDUSTRIAL AND ENGINEERING CHEMISTRY
to the solution prevents losses on evaporation, but forms precipitates insoluble in water and methyl alcohol that interfere with the quantitative extraction of boron by the alcohol. The precipitation of insoluble compounds by an alkali in the presence of considerable methyl alcohol does not present this difficulty if the alcohol is added slowly to the cold solution, By evaporating with small additions of alcoholic ammonia, approximately 90 per cent of the boron present will be easily soluble in methyl alcohol, and practically 100 per cent can be recovered by extracting for 3 hours in a Soxhlet tube. Determinations of plant material gave similar results to those obtained with salt mixtures. The recoveries from additions of boron (boric acid) to watercress are given in Table 111.
387
the sample in a muffle furnace a t or below 400' C. Cool in a desiccator, moisten the ash with distilled water, cover with a watch glass, and add a strong solution of citric acid with a finger pipet until the solution is strongly acid. Wash into the extraction flask with a small quantity of water and 30 to 35 ml. of methyl alcohol. Add 0.1 ml. of 0.5 N potassium hydroxide to 10 ml. of distilled water in the second flask, and distil for 1.5 hours. Wash the alkaline solution into
TABLE111. RECOVERY OF ADDED BORONFROM WATERCRESS (Average of three 25-gram samples) TREATMENT B ADDEDB PRESENT B FOUND Gram (from Gram 0.003 Ashed 750' F. 0:Obl 0 : b04 0.00393 Ashed 750' F. 0.00517 0.002 0.005 Ashed 750' F. 0.00798 0.008 0.005 Ashed 750' F. 0.003 0.003 Ashed 850' F. 0.00375 0:001 0.004 Ashed 850' F. None HCl soln. of residue of previous extraction . . . ... None HC1 s o h of fused silicates ... ...
To avoid the handling of large quantities of salts insoluble in methyl alcohol and the extraction of soluble salts other than borates, distillation was substituted for extraction by using the apparatus shown in Figure 3. The sample (plant ash) was dissolved in an excess of citric acid and transferred to flask C, using approximately 20 ml. of water. Washing was completed with methyl alcohol (30 to 40 ml.) and the flask connected. Flask F contains 0.1 ml. of 0.5 N potassium hydroxide in 10 ml. of distilled water. Boron is absorbed by the solution in F from the vapors produced in C. The alcohol vapor is condensed in the reflux A a n d r e t u r n e d t o flask C. The boiling point is higher in flask C and the alcohol is completely condensed, resulting in a major p o r t i o n of u the alcohol being retained G in flask C. Boron was comFIGURE3. APPARATUS FOR p letely extracted from SEPARATION OF BORONBY plant ash by this m e t h o d DIS'rILLATION in 1.5 hours. Boron could A . Condenser E. Adapter not be found in the small C. Sample flask washing tube a t the top of D. Absorption tube E . Condenser t h e c o n d e n s e r even when F . Receiver C. Water bath relatively large quantities of boron were distilled. The procedure for plant material found to be the most rapid and accurate is as follows: Weigh a sufficient quantity of the ground sample of plant material (calculated to the moisture-free condition) to give 0.0005 to 0.005 gram of boron. If the plant material yields an acid ash, add sufficient potassium carbonate to make the ash alkaline, and char in a silica dish over a low flame. Ash
FIGURE 4. APPARATUS FOR BURNING METHYL BORATEIN OXYGEN A . Cell E. E. Oxygen inlet F. C. Chimney 8 D . Microburner H: I . Alcohol
Connection for flask Vapor inlet (capillaFy) Gas inlet for ilot light Pressure reguyator overflow
a 50-ml. dish, add 5 ml. of ammonium hydroxide, and evaporate to dryness on the water bath. Dissolve the residue in 5 per cent phosphoric acid in methyl alcohol, and make up to a suitable volume with acid alcohol at 20" C. Transfer to an Erlenmeyer flask, connect to the burner, and, after the flame meets the prescribed conditions of 1.5 em., add 0.01 N potassium permanganate solution t o 50 ml. of water in the cell until the bright green band of the boron spectrum is eliminated. Refer the required quantity of potassium permanganate solution to a table showing grams of boron per liter for the quantity present, and divide by the weight of the sample to obtain the per cent of boron present (grams of boron per liter X 10/1000 X 100/sample). The table showing grams per liter of boron present for any quantity of potassium permanganate solution used should be prepared by the operator from known quantities of boron, because the sensitivity of different spectroscopes and their adjustments vary widely. Reagents used must be free from boron. The method is also applicable to the determination of boron in animal tissue and waters as well as plant tissues by slight modifications of the procedures for obtaining the boron in suitable form for solution in methyl alcohol. The method is sensitive to much smaller quantities of boron if the vapors are burned in an atmosphere of oxygen. Figure 4 shows the apparatus used for burning methyl alcoholmethyl borate vapors in oxygen under controlled conditions. The quantity of vapor burned in oxygen is greater than in air and cannot be controlled by the size of the flame, as marked changes in the rate of vaporization make but slight differences in the size of the small, hot flame. To insure the burning of the vapor a t the same rate in each determination, constant pressure was kept in the distillation flask by a mercury valve and capillary tube. The tubes D and G are of thick-
ANALYTICAL EDITION
388
,
walled hard glass tubing (not Pyrex). Attempts a t sealing these tubes (and B ) into a glass support for the cell and chimney were unsatisfactory, so a rubber stopper and a short glass collar of slightly greater diameter than the chimney 15 14
13 12
1
0
5
10
15
20
25
30
Boron mgs. per liter FIGURE 5. POTASSIUM PERMANGANATE SOLUTION REQUIRED TO ELIMINATE FIRSTGREEN BANDFROM SPECTRA OF BORONBURNED IN OXYGEN
were substituted. A metallic base supported with a ringstand clamp is recommended for the introduction of the gas tubes and support of the cell and chimney. The chimney is of heat-resistant glass. An unsteady flame due to currents of oxygen can be avoided by placing glass wool in the bottom of the chamber. Figure 5 shows the grams per liter of boron in methyl
Vol. 4, No. 4
alcohol represented by any quantity of 0.01 N potassium permanganate solution within the limits of the method. The possibility of replacing the buret and cell with a scale and sliding glass wedge similar in color to potassium permanganate solution was considered but WBS not investigated. The use of the wedge would probably simplify the absorption of the spectrum and shorten the method, but to approach the accuracy of the buret and cell, the lengths of the scale and wedge would probably be too great for convenience. The spectroscopic method by burning in oxygen will detect the presence of less than 0.5 part per million of boron in methyl alcohol. Quantities of boron between 3 and 30f parts per million can be determined quantitatively with an accuracy of 10.35 part per million, Plant materials could not be checked as accurately as could solutions of boric acid, but could be checked within the limits of *0.95 part per million, or *0.000095 per cent on a moisture-free basis. b\The method of Wilcox (IO) and the A. 0.A . C. (I) voluqetric method were not applicable to the determination of such minute quantities of boron as occur in normal plant tissues. The low buret readings on extremely large samples of plant material and the necessity of two additions of standard alkali resulted in a wide range of inaccuracy. LITERATURE CITED Assoc. Official Agr. Chem., Methods of Analysis, p. 339 (1930). Bertrand, G.,and Agulhon, H., Compt. rend., 155, 248-51 (1912). Bertrand, G., and Agulhon, H., Zbid., 156, 732-5 (1913). Bertrand, G.,and Agulhon, H., Bull. 800. chim., 7,90-9 (1894). Bertrand, G.,and Agulhon, H., Compt. rend., 157, 1433-6 (1913). Bordas, G., and Touplain, F., Ann. fals.,16,356-60 (1923). Jay, Compt. rend., 158, 357-8 (1914). Mannick, C., and Preiss, H., Chem.-Ztg., 32, 314 (1908). Robin, L.,8th Intern. Congr. Appl. Chem. 1, Orig. Com. Sec. 1, 429-32 (1912). Wilcox, L. V., IND.EXQ.CHEY., Anal. Ed., 2,358 (1930). RECBIYBDApril 8,1932. Published by permiasion of the Director, Kentucky Agricultural Experiment Station.
An Esterification Resin as a Ground-Joint Lubricant T.P. SAGER,Bureau of Standards, Washington, D. C.
I
N A paper by Bruun and Schicktanz (I), mention is made
of the use of a mixture of diethylene phthalate and triethylene phthalate resins as a stopcock lubricant for use in contact with petroleum distillates. Although this mixture was found to have the desired consistency and stability for a ground-joint lubricant, it was not adapted for use with distillates rich in the aromatic hydrocarbons, since these resins are readily soluble in benzene and toluene. It was found that a resin made with a polyglycol and a tri-reactive acid, in which the esterification was allowed to proceed only to the fusible stage, provided a lubricant having the proper viscosity which was insoluble in both the aromatic and aliphatic hydrocarbons. One mole of citric acid and 1.5 moles of tetraethylene glycol were heated together a t 180" to 185" C. for 90 minutes. The flask was then removed from the oil bath and allowed to cool to room temperature. Prolonged heating of this mixture results in an infusible product. The resin obtained was clear, amber-colored, of balsam consistency, and possessed marked adhesiveness. It was found to be readily soluble in water, alcohol, and acetone, but insoluble in petroleum ether and toluene. Chemical analysis indicated about 65 per cent esterification.
When applied by the customary technic to ground joints, it provided a satisfactory seal in contact with petroleum distillates over considerable periods of time with no tendency to become thin, and of low lubricating value. I n contrast with many lubricants suggested for this purpose, the resin offers greater stability, is chemically inert, and may be kept indefinitely without changes occurring in its properties. Tetraethylene glycol, CHZOH(CH~OCH~)~CH~OH, may be synthesized from ethylene oxide and triethylene glycol in a pressure reaction. It boils a t 190" C. a t 3-mm. pressure. The presence of the long chain separating the two reactive groups imparts fluidity when it is used as a resin base. A similar resin made with citric acid and triethylene glycol was more viscous and, when used at room temperature, presented too much resistance to turning of the stopcock. LITERATURII CITED (1) Bruun, J. H., and Schicktanz, S. T., BUT.Standards J . Research, 7, 871-2 (1931). RECEIVBD April 22, 1932. Publication approved by the Director, Bureau of Standards. ,
