V O L U M E 25, NO. 6, J U N E 1 9 5 3
917
Failure to obtain a saturated solution during preparation of the wash solution, or to bring the sample extract to the same temperature as the wash solution. Use of a poorly chosen solvent or failure to remove all solvent from the crystals. Coprecipitation of impurities with the crystals. This case has been considered above. In some cases the proposed procedure may be of assistance in the identification of impurities. An insoluble fraction in the thimble, if not dust, dirt, or filter fibers, can usually be extracted with a better solvent and recovered for examination, practically free from the major component. Soluble impurities are enriched perhaps ten- to 50-fold, depending upon the quantities employed and the solubility of the pure compound. Since the amount of known material in the soluble fraction can be calculated, one can determine the ultraviolet or infrared absorption spectrum of a solution, allow for the absorption of the known substance, and arrive a t an approximate spectrum of the impurities. Thus the
presence of one or more isomers has been indicated in some samples of p,p’-isopropylidenediphenol. LITERATURE CITED (1) Bennett, G. M., Analyst, 73, 191 (1948). (2) Butler, J. A. V., J . Gen. Phvsiol., 24, 189 (1940). (3) Cristol, S. J., Hayes, R. A., and Haller, H. L., IXD. ENG. CHEM.,ANAL.ED., 17, 470 (1945). (4)Herriot, R. M,, Chem. Revs., 30, 413 (1942). (5) hloore, S., Stein, W. H., and Bergmann, R I . , I h i d . , p. 423. (6) Northrop, J. H., and Kunitz, hl., J . Gen. Physiol., 13, 781 (1930). (7) Reeve, Wilkins, and Adams, Rowland, B N ~ LCHEM., . 22, 755 (1950). (8) Stull, D. R., ISD. ENG.CHEY.,ANAL.ED.,18,234 (1946). (9) Tarpley, William, and Yudis, Milton, ANAL CHEM.,25, 121 (1953). (10) Webb, T. J.,Ibid., 20, 100 (1948). RECEIVED for review December 13, 1952. Accepted February 13, 1953.
Construction and Calibration of Simple Semiautomatic Microburets U . L. UPSON General Electric Co., Richland, Wash. x
THE
routine performance of microtitrations, the use of screw-
1 feed mechanisms becomes tedious and time-consuming, since
no other operation can be performed during either the delivery or the refill period. I t was felt that a microburet offering the convenience and familiarity of stopcock control would be highly desirable. TKOsuch models were developed, covering the microand ultramicro ranges, and are described here. A method for the calibration of all Rehberg-type microburets which involves refinements yielding increased precision over previously reported methods is also given. AIR-CONTROLLED MlCROBURET
For microtitrations involving 200 pl. to 2 ml. of titrant, especially those of a repeated or routine nature, the easily constructed apparatus described here has been found equally precise and considerably more time-saving than the commercially available models employing the Rehberg-type control (2,3,4). Delivery is controlled by manipulation of a stopcock, just as for an ordinary semimicroburet, using the flip technique of incremental addition near the end point, and filling is accomplished semiautomatically. The device employed is not practical for increments of less than about 0.2 pl. but has been found highly satisfactory for capillaries of from 0.6 to 2.0 mm. inner diameter, giving capacities of from 200 pl. to 2 ml. for 25 to 35 cm. scale length. The delivery is controlled by regulating the flow of the air displacing the titrant in
the buret tube, and the air flow is throttled by means of a small piece of capillary tubing cemented into the bore of the stopcock. This tubing should be somewhat less than the bore length to prevent clogging with grease and should have an inner diameter of from 0.005 to 0.010 inches, depending upon the buret capacity, and hence permissible increment, chosen. Since the flow rate varies both with this throttling and with the hydrostatic head, the slope of the tube can be somewhat greater, or the tip longer, for the larger capacity burets than for those of less than 500 pl. volume. The buret tip should be drawn down and ground back to limit the flow, with the stopcock removed, t o about twice the desired maximum delivery rate, but in order to ensure equilibrium pressure and thus to retain control without overshoot, the bore of the stopcock plug must remain the ultimate flow rate limiting factor. For a maximum rate of about 4 cm. per minute, the precision is limited only by the reading error. To prevent dust from plugging the throttling capillary or contaminating the buret, a filter is attached to the air input tube. This can be a loose wad of long staple cotton in a drying tube (flushed with air before using). ADJUST
TO TO BURET
FILTERED ( I ATMOY
I i Figure 1.
Air-Controlled Microburet
For capacities from 200 PI. t o 2000 PI.
Figure 2.
Device for Adjusting Vacuum
The original model (Figure 1) is filled by immersing the buret tip in a bottle or beaker containing the titrant and by turning the stopcock to apply a slight vacuum. This vacuum is adjusted (a device such as that shown in Figure 2 is suggested) so that the
978
A N A L Y T I C A L CHEMISTRY
titrant rises in the buret to a height a t least as far above zero scale as the tip is immersed. The stopcock is then closed, the Supply removed, and the meniscus leveled a t zero scale by quickly flipping the stopcock back and fourth past the air-input position. A modification especially useful for routine analyses employs a two-way stopcock near the buret tip and a titrant reservoir whose height can be adjusted to refill the buret to the desired level. In this case, of course, no vacuum is needed. Burets of this type generally exhibit appreciable variation in bore along the capillary length, and care should be taken to select the tubing for minimum variation along the scale portion. The buret is best calibrated, a t 1-cm. scale increments, by the mercury segment method described below. OILCONTROLLED ULTRAMICROBURET
For microtitrations involving 20 to 500 p l . of titrant, a modification of the above buret has been designed nhich affords comparable precision and ease of operation.
II
DELE i' R
The oil is introduced by means of a leveling bottle through the drain cock, Cs, and is brought to the level A in both arms. The titrant is drawn into the buret, from a bottle into which the buret tip is immersed, by turning the upper stopcock, C,, to the vacuum and by opening the lower stopcock, C2 The lower stopcock, Cp, is then closed, the upper stopcock, C1 is reversed, and the meniscus is adjusted to zero scale by means of the lower stopcock, C p , using the flip method of incremental delivery. Delivery is effected as a result of the slight hydrostatic pressure of the oil. The vacuum must be adjusted so as to draw the titrant to a level betxr-een zero on the buret scale and D,but once the vacuum has been properly adjusted, the titrant automatically will be brought to the desired level for each filling, without further adjustment. The bore of the tube between A and B should be sufficient to allow the ~a11sto drain properly and to avoid the formation of air bubbles which might carry oil into the capillary section, 6 mm. inner diameter being sufficient. The bulbs E and F should be equal to the capacity of the largest buret tube to be used with the control. The entire assembly, control, scale, and buret tube, should be rigidly mounted on a panel for strength and convenience. The control device may be fitted u-ith a standard taper ground joint and interchangeable capillary buret tubes of various capacities may be attached, as indicated, or a single capillary tube may be incorporated as an integral part of the buret, as in the aircontrolled model (Figure 1). If different capillary tubes are used, a zero reference line should be scored on each, and the zero scale should be adjusted to it for each interchange. If this is done, each tube need be calibrated only once. CALIBRATION OF MICROBURETS
WTERCHANGEABLE CAF'ILLARY BURET
Figure 3.
B
Oil Control for Ultramicroburet for Capacity from 20 pl. to 500 pl.
In the control apparatus (Figure 3), the titrant f l o is ~ regulated by means of a capillary-bore stopcock, Cz, through which oil flows under slight pressure. The rate of flow, and hence the minimum volumetric increment, varies with both the viscosity of the oil and the size of the capillary. By using a light weight vacuum pump oil (e.g., 2 parts HyVac to 1 part mineral oil) and by having three interchangeable stopcock plugs of different bore [made by cementing in capillary tubes of from 0.6 mm. to about 2 mm. inner diameter ( I ) ] , the same control can be used over a wide range of capacities, with &0.1% reproducibility and with desirable drainage rates. Here, as in the air-controlled model, the buret tip should be adjusted to restrict the flow to something greater than permitted by the bore of the flop--controlling stopcock. A maximum flow of less than 5 cm. per minute yields maximum precision.
The most satisfactory method for the calibration of micro- and ultramicroburets of the types in which a capillary meniscus is read against a uniform linear scale has been found to be that used by Benedetti-Pichler ( 1 ) and by Kirk ( 3 ) ,employing a short mercury segment to determine variation in the cross sectional area. In this method a small quantity of mercury is drawn into the capillary tube, the volume being carefully adjusted to give a segment of very nearly unit scale length. This mercury segment is, then, a small cylinder of constant volume (temperature being held constant) whose length varies inversely with the average cross sectional area of the included tube segment. If the precise length of this segment is measured for each major scale increment (e.g., 1 to 2, 2 to 3 cm., etc.), the ratio of this length to the average for the entire scale varies inversely with and is a measure of the ratio of the corresponding incremental volume (for one scale division) to the average volume per unit length. This last value is obtainable by weighing the total water for full scale delivery and dividing the resultant volume by the total scale length. (In this laboratory, and in the evample given here, volumes are corrected to a reference temperature of 25' C.) The procedure previously used in this laboratory was that recently described by Kirk ( 3 ) . Because of the nonuniformity of most commercial burets and stock capillary tubing, several modifications of Kirks technique were introduced to effect greater calibration accuracy. One modification was to restrict more closely the permissible range for the segment length. In this calibration the variable length of a constant volume of mercury is measured to determine the variable volume of a constant incremental length of capillary tubing. It is a desirable condition, then, that the volume evaluated be as nearly as possible identical to that included in the nominal scale portion. To this end, the mercury segment should be so taken that it does not differ from the scale unit length (usually 1 cm.) by more than 5% (0.5 mm.) a t any position on the scale. While the errors induced by a greater discrepancy may tend to average out, they may become cumulative for capillaries having appreciable taper, and in such
979
V O L U M E 2 5 , NO. 6, J U N E 1 9 5 3 Table I. Scale, Cm. 0-1 1-2 2-3 3-4 4-5 , , , d
31-32 32-33 33-34 34-35
Calibration of Microburet
Li
Lz
1.02 1.02 1.02 1.01 1.01
1.02 1.02 1.02 1.01 1.01
La 1.02 1.02 1.02 1.02 1.01
Lau. 1.020 1.020 1,020 1.013 1.010
Va 3.063 3.053 3.053 3.074 3.083
1.00 1.00 1.00 1.00
1.00 1.00 1.00 0.99
1.00 1.00 1.00 0.99
1:OOo
0.993
3:ii4 3.114 8.114 3.136
ZL
35,227
...
...
...
=
1.000 1.000
.ZV Factorb 3.053 3.053 6.106 3.053 9,159 3.053 12.223 3.055 15.316 3.063
....
98.925 102.039 105.153 108.289
0.10821 23.5 1.00366 108.61 35.10 3.0943 A v
3:oQi 3.091 3.092 3.094
=
3.0940
V = K/La,. ZV/scale cm. True capacity a t 25’ C. per gram of water a t 1’ C. d D a t a omitted.
a
b
C
cases the segment length should be carefully adjusted to straddle the scale unit length. I n the former procedure, the per cent volumetric deviation for each scale increment is taken to be equal to the corresponding per cent deviation in segment length, but of opposite sign. Khile this approximation introduces no significant error for capillaries having only small deviations-Le., under 2’3-it may introduce errors exceeding 0.2% if the deviations are as great as 5y0. The procedure described below tends to minimize this error and is somewhat easier to apply. The procedure given here reduces all methodical error to l e v than 0.1% for capillaries having up to 10% total variation in inside diameter. I n practice, no discrepancies of greater than 0.03% have been found using the procedure given here. PROCEDURE
Clean the capillary 171th concentrated nitric acid and rinse well with distilled water.
For the incremental calibration draw clean mercury into the buret (from a porcelain microtitration dish) to give a segment length of one major scale division (1 cm.) within 5y0 a t all divisions. iidjust position of segment to cover the 0 to 1 division, estimate length to nearest 0.01 cm., and record ( L ) . (See Table I. j Repeat this step for each major division on the scale. Obtain two more sets of readings ( L zand L B )and , average (Lav,). For the total volume determination fill the buret with distilled water a t room temperature and record the temperature. Deliver full scale volume into a tared, dry, 5-ml. volumetric flask. Stopper immediately. ( T h e n zeroing buret and delivering to flask, touch off the drop on the tip. Deliver a t a slow, even rate.) Record difference between initial and final scale readings. Determine net weight of delivered volume and convert to true capacity in glass ( a t 25” C.) in p l . Divide by scale reading to give pl. per cm. Repeat these steps to give three values, and record average p l . per cm. For the calculations, set up the figures as shown in Table I. Determine average segment length by dividing the sum of the Lau, values (ZL) by the number of scale divisions. Determine the calibration constant, K , by multiplying the average p l . per em. value by the average segment length. Calculate the incremental volume, 8, for each scale division by dividing k’ by the La”,value for that division. Calculate the summation of the volumes from zero to each major scale division. (The summation for full scale must equal the true capacity volume.) Calculate calibration factors by dividing each summation volume by the corresponding scale reading. These factors may be plotted graphically for convenience in determining intermediate values. Volume (PI.) = scale reading (cm.) X calibration factor. LITERATURE CITED
(1) Benedetti-Pichler, A. -4., “Introduction to the Microtechnique of Inorganic dnalysis,” p. 258, New York, John Wiley and
Sons, 1942. (2) Cunningham, Burris, Kirk, P. L., and Brooks, S. C., J . Biol. Chem., 139, 11-19 (1941). (3) Kirk, P. L.,“Quantitative Ultramicro Analysis,” p. 36, Sew York John Wiley & Sons (1950). (4)Sisco, R. C., Cunningham, Burris, and Kirk, P. L., J . B i d . Chem., 139, 1-10 (1941). R E C E I T E for D review .ipril 7, 1952. Accepted February 9, 1953. Presented before the Sorthwest Regional .\feetin&? of the .%MERIc.As CHEMICALSoCIETY, Corvallis, Ore., .June 1952.
Measuring the Resistance of Polarographic Cell Circuits RI. R. PESCE, S. L. KNESBACH, AXDR. K. LADISCH Pioneering Research Laboratories, U . S . Army Quartermaster Corps, Philadelphia 45, Pa.
P
o r d u z o G R I r w c circuit re-istancei can be measured accurately and conveniently by the use of an oscilloscope as
the null point detector in a Wheatstone bridge arrangement. Ordinarily, the polarographic half-wave voltage read from a c.-u. curve must be corrected for iR acrosls the circuit in order to arrive at the characteristic half-wave potential of the test substance. The current, 2, is readily obtained from the curve, while the measurement of the resistance R is more difficult. Ilkovii: (3)introduced for the latter purpwe the Kohlrausch method emploring an alternating audible signal of unspecified frequency. Muller (8)prefers a method of approximation, stepTvise increasing the applied e.m.f. to a high value and calculating R by means of Ohm’s law. In thi- manner, the voltage lost by polarization a t the dropping electlode becomes negligible. Elsewhere ( 7 ) , he mentions a number of other methods. For instance, he utilizes the shift in half-wave potential of a given substance on a welldefined curve, resulting from variations in wave height, to calculate R by means of Ohm’s law Furman and Bricker ( 1 ) measured R with a conductivity bridge containing an electron ray tube balancing device, a t a frequency of 1000 cycles per second. It is claimed also that R may be determined in certain special cases from the slope d R / d , (2). These methods are not very precise, and some of them are insufficient even for a reasonable estimate of the circuit resistance.
Errors aiiqe, for inqtance, whenever the E/z relationship is used to calculate R from small incrementq or differences in potential and current, the size of which may actually approach the limits of experimental error. I n other cases, the resistance is determined a t applied voltages far apart from the half-wave potential. This changes the drop time with a concurrent change in resistance ( 3 ) . Until recently, a precision of a 1 mv. per pa. in the half-wave potential or of 500 to 1000 ohms in the resistance has been considered adequate (4,5 , 8). 1lkovii:’s method is suited for this purpose, although it is somewhat cumbersome to use. However, caution must be exercised in more precise polarographic studies such as reported by hleites ( 6 ) , who aims a t half-wave potentials with a probable error of 5 0 . 2 mv. a t 2.5 pa. For refined determinations like these, the bridge method can be improved considerably by introducing an oscilloscope as the null point detector. We have used a 5-inch DuMont 208B oscilloscope having a 10 mv. per inch deflection sensitivity (vertical amplifier), in combination with a Leeds and Northrup #4725 bridge. Equally satisfactory results were obtained with 3 inch oscilloscopic equipment, using an external amplifier to obtain the desired sensitivity. I n operation, the polarographic circuit is made part of the bridge and a sine wave signal of the desired frequency is fed into the system. The bridge is then balanced, using the minimum vertical deflection of the oscilloscope