Efficient Vacuum Fractional-Sublimation Apparatus Keith Gosling and Richard E. Bowen Department of Chemistry, West Virginia University, Morgantown, W. Va. 26506
In the course of research involving an air-sensitive aluminum-nitrogen ring compound, it became evident from NMR data that we were dealing with a mixture of geometric isomers. Isolation of the individual isomers by fractional crystallization was precluded by quite rapid isomerization in solution and conventional vacuum sublimation (1, 2 ) on to a cold-finger proved to be ineffective. However, vacuum sublimation up a 200-mm long Pyrex tube surrounded by a close-fitting copper jacket did produce partial separation, if the temperature of the oil bath heating the lower 20 mm of the tube was carefully controlled. Overlap of the fractions, which could be recognized by the different crystal habits of the isomers, indicated that separation employing a longer tube, with a smaller temperature gradient, would be more efficient. The vacuum fractional-sublimation apparatus described below embodies significant improvements over the gradient sublimer used by Melhuish ( 3 ) in that it permits the introduction of airsensitive compounds without exposure to air and provides a variable temperature gradient facilitating separation of compounds which differ only slightly in their vapor pres-
to vacuum line t o nitrogen supply
r.
a i r out
z
-vacuum
!
jacket
I
r
ampouh holder
.
sublimation t u b
a i r in
J
asbestos papzr wedges
flowmeter kutor
t o vartac Figure 1.
Construction of vacuum fractional-sublimation appara-
tus (1) K. B. Wiberg, "Laboratory Technique in Organic Chemistry." McGraw-Hill Book Company, Inc.. New York, N.Y., 1960, p 114. (2) R. J. McCarter. Rev. Sci. Instrum., 33, 388 (1962). (3) W . H . Melhuish, Nature, (London). 184, 1933 (1959).
1574
ANALYTICAL C H E M I S T R Y , VOL. 45, NO. 8, JULY 1973
sures. This apparatus has been most effective in the separation of closely similar isomers and has proved equally useful for the final purification of other sublimable compounds. The apparatus (Figure 1) consists of a 1000-mm long, 11-mm 0.d. Pyrex sublimation tube which is sealed a t the lower end and a small indentation created to facilitate the breaking of the sample ampoule. The upper end is equipped with a 14/20 standard-taper outer joint for attachment to a vacuum line via a right angled elbow. This allows the sublimation tube to be pivoted outward for easy assembly where vertical space is limited. The vacuum line carries a 3-way stopcock a t this point to permit evacuation of the sublimation tube or admission of dry nitrogen. A 900-mm long, 15-mm i.d. unsilvered vacuum jacket surrounds the sublimation tube and is centralized by three, 3-mm wide, asbestos paper wedges a t top and bottom. The jacket is clamped independently to the supporting framework. The assembly is completed by a nichrome heater coil enveloped in a Pyrex tube and attached to the lower end of the vacuum jacket by a 28/15 spherical ground joint. In operation, compressed air from the house supply is led via a flow-meter, to the heater tube, then u p the annular space between the sublimation tube and the vacuum jacket and exhausted a t the top. We have found that the compressed air supply which is regulated to 20 lb/in.2 in the main manifold, supplies a sufficient volume of air a t virtually constant pressure to provide a temperature drop of 3 "C over the entire length of the jacketed tube a t an operating temperature of 100 "C. A lower air velocity increases the temperature drop. Sample ampoules (Figure 1, inset) are made from 8-mm 0.d. Pyrex tubing with a 6-mm 0.d. neck. The lower end is sealed and blown out to form a thin-walled bulb. For airsensitive compounds, the ampoules are loaded in a drybox and sealed a t the upper end. Insertion of the sample into the sublimation tube is accomplished with the ampoule-holder (Figure 1, inset) which is fabricated from a 20-mm length of thin-walled stainless steel tubing brazed to a 1200-mm brass rod. The tubing is slit longitudinally in two places to provide a clamping action on the glass ampoule. A typical procedure for the vacuum sublimation of an air-sensitive compound is given below and a simplified operation for air-stable samples will be evident. Experimental Operation. A sublimation tube was flamed out under vacuum, filled with nitrogen, and removed from the vacuum line. The sample ampoule was introduced with the holder, broken on the protrusion a t the bottom of the tube, and the sample shaken out. With the ampoule-holder removed, the sublimation tube was attached to the vacuum line and pumped out immediately. (Very fine particles had a tendency to "float" in the tube when vacuum was first applied or when the temperature was increased. A fine glass-wool plug inserted directly on top of the crude material in the sublimation tube prevented this from being a problem.) The vacuum jacket and heater tube were fitted and connections made to the air supply and the Variac. The air supply was turned on fully to give a constant reading on the flow-meter and the
temperature of the air stream was increased gradually by means of the Variac control until sublimation began. Further slight increase in temperature caused movement of the sublimate further up the tube. When sublimation was complete, the heater was turned off and the air supply interrupted. The heater tube and vacuum jacket were removed and nitrogen was introduced into the sublimation tube to within 5 mm of atmospheric
pressure. The tube was sealed off with a torch a t convenient places to include separate samples for transfer to a dry-box. Received for review November 30, 1972. Accepted March 14, 1973. Presented in part as paper No. 152, Inorganic Chemistry Division, 163rd National Meeting of the American Chemical Society, Boston, Mass., April 1972.
Simple, Efficient Micro Mixing Device Robert E. Gugger and Samuel M. Mozersky Eastern Regional Research Laboratory, Agricultural Research Service, U.S. Department of Agriculture, Philadelphia, Pa. 797 18
Thorough mixing of two streams of solution, A and B (Figure l), is important in many kinetic measurements and in continuous assay procedures. Such mixing can be easily accomplished with high velocity Get) streams, i.e., those having velocities of the order of 100 cm/sec or more, where turbulence is readily achieved, as shown by Hartridge and Roughton ( 1 ) . For low velocity streams, i.e., those moving ca. 10 cm/sec or less, the combined solution C is frequently passed through a mixer M consisting of a coil of tubing. However, it is necessary that the time interval for flow of successive portions of solution from the point of mixing, Y, to the point of observation, 0, be constant. To ensure that this constancy is not destroyed during the mixing process, the stream is frequently segmented by bubbles of air prior to entering M. The mixing device described here required neither a mixing coil nor air, nor the peristaltic pump frequently used in such work. The mixing chamber consists simply of a Teflon (DuPont)-coated bar magnet about 2 mm in diameter x 7 mm long enclosed in a glass tube ca. 4-mm i.d. (Figure 2). The bar is restricted to a section of tubing about 10 mm long. This can be accomplished by reducing the diameter of the glass tubing a t one end, and inserting the bar magnet through the other end. An 8-mm length of tightly fitting plastic tubing is then inserted into the wider end, to keep the magnet in place. The wider end of the glass tubing cannot be heated to reduce its diameter since this might melt the Teflon coating on the stirring bar. When the mixing chamber is placed on a magnetic stirrer, the primary motion obtained is rotation of the bar about its cylindrical axis. In addition, the bar wobbles, and there is some movement to and fro in the direction of the axis of the tubing, both of which enhance the stirring action. End-over-end rotation about a vertical axis, the usual motion of a stirring bar, is, of course, impossible, since the i.d. of the glass tubing is less than the length of the bar. The mixing action is easily controlled by adjusting the rate of rotation of the driving magnet in the magnetic stirrer. The volume of liquid in the central 10-mm section of the mixer is ca. 0.1 ml. There is thus little opportunity for unwanted mixing of portions of liquid of different age (age 0 being the instant of combination at point Y, Figure 1). The efficiency of the mixing device was tested as follows. Two Y tubes made from 2.5 mm (internal diameter) (1) H . Hartridge and 376 ( 1923).
E. J.
W. Roughton,
Roc. Roy. SOC., Ser.
A , 104,
Y
M
0
B/ Figure 1. Arrangement for continuous assay procedures Y is the point of confluence of the two streams, A and B, which contain the reactants. The resultant stream C flows through a mixer M before passing the point of observation 0
I 4mm. I P l a sI t i c Tubing
'
r 7 m m , 1
I
I
Stir'ring B a r
Figure 2. Micro mixing device. The direction of liquid flow is from right to left
Magnetic Stirrer
Figure 3. Experimental arrangement for testing the efficiency of mixing
Pyrex tubing were jointed either directly, by a piece of Tygon tubing, or via the mixing device, as shown (Figure 3). The arms of the Y tubes were ca. 1.5 cm in length. The Y tubes were positioned in a common vertical plane, with the connecting tube horizontal. When a mixer was used, a magnetic stirrer was placed below it. A solution of 0.1M NaOH containing sufficient phenolphthalein to give ANALYTICAL CHEMISTRY, VOL. 45, NO. 8, JULY 1973
1575
