Variable Gradient Device for Chromatography - Analytical Chemistry

---

Variable Gradient Device for Chromatography E. A. PETERSON and

H. A. SOBER

laboratory of Biochemistry, National Cancer Institute, National Institutes of Health, Public Health Service, U. S. Department o f Health, Education, and Welfare, Befhesda, Md.

b The gradient devices described in the literature do not offer the flexibility in adjustment that is required in many A chromatographic applications. mixer that can produce nearly any gradient without change in the apparatus consists of a series of identical mixing chambers which contain eluent of varying composition in hydrostatic equilibrium. The initial concentration of each eluent and its position in the series are the controlling factors in the shape of the gradient. Quantitative relationships, as well as construction details, are presented.

S

the introduction of gradient elution t o chromatography several years ago, there have been many applications of the general technique in the fractionation of mixtures of substances having a wide range of elution requirements, and several methods for the production of gradients have appeared in the literature (1-13). With the exception of those which can utilize differential pumping rates (9, I $ ) , all require a change in a t least a part of the apparatus for each change in the shape of the gradient. The need for a convenient means of making m a l l changes in specific portions of the elution gradient in order to improve resolution in certain rpgions of the chromatogram, as well as for presenting the gradient data in a manner to permit ready duplication by others, led to the development of a variable gradient mixer (T'arigrad) made up of a number of identical chambers connrcted in series and arrtinged in compact form. INCE

I

Two arrangements employed in this laboratory are shown in Figure 1. I n use, all the chambers are in hydrostatic equilibrium which is constantly being re-established in response to the removal of liquid from a chamber a t one end of the series. -4 central motor drives a simple mechanism which stirs the liquid in each chamber. It is obvious that this system is part of a homologous series, of which Parr's two-chambered apparatus (11) is the first member. Because the authors' first model was of the square type shown in Figure 1, a nine-chambered system was a logical choice, and it has proved satisfactory in both mechanical performance and flexibility in the production of gradients. SIMPLE GRADIENTS

Any number of the nine chambers can be used in combination by closing the appropriate connecting channel. If solution of concentration L (representing the limit concentration) is placed in the last chamber of a given combination, and the others are filled cvith water, liquid emerging from the first chamber will change in composition according to the curves shown in Figure 2. With L in chamber 2 of a two-chambered system, a linear gradient will be obtained. K i t h L in chamber 3 of a three-chambered system, a concave gradient results, and the curvature increases progressively as the number of chambers is increased. The concentration, C, emerging from the

where L' is the volume of liquid which has emerged up to that point, V is the total volume of liquid originally in the mixer, and h' is the number of chambers in the system. COMPOUND GRADIENTS

If solution of concentration L is placed in a chamber which is not the first or the last of a system, curves of an entirely different sort are obtained. This is demonstrated in Figure 3, which presents all the curves resulting from single-chamber contributions in a nine-chambered series. Those produced by placing L in chamber 2 or in chamber 8 are mirror images of each other; the same is true for 3 and 7. and 4 and 6. 4 . symmetrical curve is obtained with L in chamber 5. These gradients can be regarded as the contributions of individual chambers to the compound gradients which the mixer is intended to produce. They can be determined empirically or calculated by the following equation: C

( X - l)!

z = (S - n ) ! ( n- l)! ;)'-" ($-'

(2)

in which S is the total number of chambers in the system, and n is the number of the chamber which contains

8 - 7 9

Figure 1.

mixer a t any point is given by:

Two arrangements of mixing chambers,

showing direction of flow

FRACTION OF TOTAL VOLUME, '/V

Figure 2. Gradients produced b y several systems with limit solution, I, in end chamber and water in the rest Number of chambers in each system is indicated

VOL. 31, NO. 5, MAY 1959

e

857

GRADIENT COMPOhENTS

9-Chambered System

Figure 4. Formation of a compound gradient b y the summation of the contributions of four chambers in a nine-chambered system F R A C T I O V OF TOTAL VOLUME, V/V

Figure 3. Single-chamber chambered system

contributions

in

nine-

Numbers within circles at top represent concentrations, relative to 1, charged to corresponding chambers. Other chambers contain water. Numbered curves represent resulting contributions of individual chambers indicated. Heavy curve i s sum of curves below it.

Blacked-in circles represent Ihe position of solution L, other chambers ore filled with water

L. The numbering begins a t the chamber from which the liquid is withdrawn. Table I presents calculated values for the contributions of individual chambers in nine-, eight-, seven-, six-, five-, four-, and three-chambered systems. Some idea of the effect of using larger numbers of chambers can be obtained very easily by sketching out the shape of the symmetrical curve which is always produced when L is in the middle chamber of an odd-numbered series. The sharpness of such a curve, as measured by the ratio of the concentration emerging a t v/V = 0.5 to that emerging a t v / V = 0.25, is given by: (3)

For a nine-chambered system (iV - n = 4) this ratio is 3.16; for a 17-chambered system ( N - n = 8) it would be 10. The actual concentration a t the mid-point, relative to L, is the product of the following series: ( N - 2) 2 ' 43 . 65 . . . . . . . .(F=q

(4)

Whereas the shape ratio increases exponentially with increasing number of chambers, the maximum height decreases by a diminishing factor. If values are required for the maxima of the unsymmetrical curves, Equation 2 must be used, but the calculation is somewhat simplified by knowledge of the positions of the maxima: (5)

The summation of the contributions of four chambers in B nine-chambered system to form a compound gradient is il-

858

ANALYTICAL CHEMISTRY

Figure 5. Illustration of a method for matching a preselected gradient, represented b y the heavy line

04 FRACTION

lustrated in Figure 4. I n this case, 0.2 of the limit concentration was placed in chamber 3,O.S in chamber 7, and'the full limit concentration in 5 and 9, with water in all the rest. I n this summation it would not be necessary to regard a starting concentration above zero as a base line to be subtracted from the limit, as the chambers make their contributions independently. The starting concentration would then be the concentration placed in chamber 1, and its contribution would be calculated in the usual way. I n chromatographic applications the shape of the required gradient is usually determined by a succession of trials. Therefore, convenience in calculating the exact curve is less important than convenience in adjusting the shape to tit requirements indicated by previous chromatograms. Such adjustment can be accomplished by increasing or decreasing the concentrations in individual chambers, and when a suitable gradient has been achieved, it can be recorded by simply listing the concentrations used. These can be presented as molar values, but it is convenient to express them as fractions or percentages of the limit con-

06

08

10

OF T O T A L VOLUME, V/V

centration. Thus, for example, 0, 2, 2, 8,8,8,15,30,and 100 define a particular shape in general terms, and a particular gradient is completely defined when the limit concentration is also specitied. For graphical presentation the gradient can be calculated from the concentration data, or determined empirically. Picric acid is a convenient solute for this purpose, because its absorbance at 355 mp is a linear function of concentration in the range which can be read in a spectrophotometer. When a mixture is being chromatographed for the first time, it is convenient to use a linear gradient with a limit concentration well beyond the anticipated maximum elution requirement. Such a gradient can, of course, be obtained by using only two of the chambers. For larger volumes a similar result can be achieved in a multichambered system by distributing the chamber concentrations in a linear manner, with the starting concentration in chamber 1. If the chromatogram resulting from the use of a linear gradient thus obtained is crowded in some portions and spread out excessively in others, the concentrations in the individual chambers which

Table 1.

v/ v 5 0.00

0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.00

V/V" 0.00

0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.00

Calculated Contributions of Individual Chambers in Systems Comprising Several Chambers

3 Chambers 3b 2 0,000 0.000 0,095 0.002 0,010 0.180 0.022 0.255 0.320 0.040 0.375 0.062 0.420 0.090 0.455 0.122 0,480 0.160 0.202 0.495 0,500 0.250 0.495 0.302 0.480 0,360 0 422 0.455 0.490 0.420 0.375 0 562 0.640 0.320 0.722 0.255 0.810 0.180 0.902 0.095 1,000

0.000

1b

2

ib

7 Chambers 6 5

4 Chambers 4 3 0.000 0.000 0.000 0.007 0.001 0.027 0.003 0.057 0.008 0.096 0.016 0.141 0.027 0.189 0.043 0.239 0.064 0.288 0.091 0.334 0.125 0.375 0.166 0.408 0.216 0.432 0.275 0.444 0,343 0.441 0,422 0.422 0.384 0.512 0.325 0.614 0.729 0.243 0.135 0.857 1.000 0.000 1 2

5 0.000

5 Chambers 4

0.000 0.000 0.000 0.002

0.004 0,008 0.015

0.026 0.041 0.062 0.092 0.130 0.178 0.240 0.316 0.410 0.522 0.656 0.814 1.000 1

0 000 0 000 0 004 0 012 0 026 0 047 0 076 0 112 0 154 0 200 0 250 0 300 0 346 0 384 0 412 0 422

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.002 0.000 0.000 0.001 0.015 0 000 0 000 0 006 0 042 0 000 0 002 0 015 0 082 0 000 0 004 0 033 0 132 0.001 0,010 0.060 0.185 0.002 0.020 0.09.5 0.236 0 004 0 037 0 138 0 276 0 008 0 061 0 186 0 303 0 016 0 094 0 234 0 312 0 028 0 136 0 278 0 303 0.047 0.187 0.311 0.276

8 0.000 0.000 0.000 0.000 0.000 0 000 0.000 0 001 0.002 0 004 0.008 0 01.5

7 0 0 0 0 0 0

000 000 000 000 000 001

6

0 0 0 0 0.028 0 0,049 0 0.082 0

017 0.077

014 049 098 154 211 265 310 346 368 375 368 346 310 265 211 154

0.000 0.000 0.000

0.001

0,002

049 0.014

0 000 2

0 000

3

1

a

0.029 0 058 0 097 0 144 0 194 0 239 0 273 0 293 0 290 0.268 0 327

098

-

9

0.000 0 000 0 000 0 000 0 000 0 000 0 000 0 000

0.001 0 002 0.004 0 008 0 017 0 032 0 058

9 8 0.000 0 000 0 000 0 000 0.000 0 000 0 001 0 003 0 008 0 016 0 031 0 055 0 090 0 137 0 198 0 266 0 336 0 385 0 383 0 279 0 000

0,008

0.000 0 002 0 006 0 015 0 028 0 049 0 077 0 113 0 156 0 206 0 259 0 312 0 360 0 306 0 410 0 392 0 328 0 204 0 000

0.024 0.051 0.088 0.132 0.181 0.230 0,276 0.312 0.337 0.346 0.336 0 309 0.264 0 205 0.138 0.073 0.021

2

3

Chambers 7 6 0 000 0 000 0 000 0.000 0 001 0 003 0 010 0 022 0 041 0 070 0 109

0 0 0 0 0 0

0.000

1.00

0.95 0.90 0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10

0.0.5 0.00 v/

va

5

000 0 000 000 0 000 000 0 005 003 0 018 009 0 046 023 0 086

0.047 0 136 0 081 0.187 0 123 0.233 0.171 0 263 0 218 0 273 0 257 0 263 0 279 0 233 0 279 0 187 0 254 0 136 0 207 0 086 0 147 0 046 0 084 0 018 0 031 0.005 0 005 0 000 0 000 0 000 4 5

1.00

0.95 0.90 0.85 0.80 0.75 0.70

0.65 0.60 0.55 0.50 0.45

0.117 0.164 0.157 0.214 0.40 0 209 0.261 0.3.5 0.075 0.244 0.328 0.236 0 259 0.298 0.30 0.118 0.303 0.324 0.185 0 297 0.318 0.25 0.312 0 178 0 356 0 297 0 132 0 100 0.311 0 173 0.20 0 262 0 393 0 246 0 082 0 168 0 293 0.275 0 115 0.15 0 377 0 399 0 176 0 042 0 272 0 238 0.210 0 062 0.10 0 531 0 354 0 098 0 015 0 149 0 475 0 3 i 2 0.124 0 023 0 430 0.05 0 735 0.232 0 030 0 002 0,698 0 257 0.041 0 004 0 663 0 OC5l 0.00 1.000 0.000 0.000 0.000 0 000 1 000 0 000 0.000 0 000 1 000 l b !i 3 4 1 2 v/va 3 4 1 2 3 a Scale a t left is to be used for chamber numbers appearing a t top of column, that at right for rhamber numbers appearing at bottom. Each column thus describes a pair of mirror images or, in some cases, a svmmetrical curve. Numbers ahove or below columns refer to chamber initially containing solution having a concentration of unity. Initial concentration in other chambers of a given series is zero.

dominate those portions of the chromatogram can be decreased or increased t o provide a more favorable distribution of eluting power in the next experiment. I n cases where the early portions are crowded and require more or less uniform expansion, xhereas the later peaks need sharpening, i t is convenient to retain the linear form of the gradient in the first part, but with a lower slope, and swing u p to the original limit in the later portions. This is readily achieved by reducing the concentrations of all the chambers except the first and last by some constant factor. A method for matching the shape of a preselected gradient, using a nine-chambered mixer, is shown in Figure 5. A thin straight line has been drawn to approximate the major portion of the desired gradient. The intercept of this line on the ordinate at the right establishes a

032 055 087 130 185 248 0.133 0.312 0.210 0 367 0.321 0 396

000

0.410 0 368 0 292 0 172

0.000 0.000 0.000 0.000 0.000 0 003 0,001 0 011

0.004 0.012 0 004 0.025 0 008 0.047

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

6 Chambers 5 4 0,000 0.000 0 000 0,001 0.000 0.000 6

0,005 0.010 0.018 0.031 0.050 0.078 0.116 0.168 0.237 0.328 0,444 0.590 0.774 1.000

8 Chambers

4

3

tentative limit from which chamber concentrations can be calculated for the production of the gradient represented by the thin straight line in the manner described above. B y substituting the limit concentration of the desired curve in chamber 9 and adjusting the concentrations of chambers 3 and 6 to provide the required increment and decrement at single points on the curve dominated by those chambers, a set of chamber concentrations is obtained which mill produce the gradient shown by the open circles. I n estimating the decrease in the concentration in chamber 6 that is required to lower the gradient to the heavy curve, one must also take into account the effect, a t that point, of the large concentration increase in chamber 9. The point selected is that a t n-hich the chamber (in this case, 6) makes its maximum

contribution, and the rcquircd cliniige is calculated from

where m is the ninsimum contribution provided by concentration C, and Am is the total change desired in this contribution. GENERAL CHARACTERISTICS

The shape of the normalized gradient is independent of the total volume, permitting easy scaling u p or down of chromatographic experiments, often with the same mixer. If desired, a single apparatus can be used to supply identical gradients simultaneously to several columns by pumping liquid from chamber 1 to all the columns at equal rates. Any number of independent gradients of different molecular or ionic species can VOL. 31, NO. 5 , MAY 1959

859

-

Y

M0

Figure 6.

Nine-chambered Varigrad

be produced simultaneously. For example, gradients of the acidic and basic components of buffers can be developed which satisfy independently selected specifications for pH and total molarity, and a gradient of a nem- buffer species can be used to compensate for the diminishing effectiveness of an initial species as the pH of the total gradient changes. Similarly, protective molecules can be caused to increase in concentration as some critical region (of pH, ionic strength, etc.) is approached. Unless prolonged interruption of stirring occurs accidentally during an experiment, the gradients produced by this type of apparatus will be free of abrupt changes in composition, such as might occur in systems in which mixing of two or more streams of liquid is effected in a small volume. Moreover, because geometry is the determining factor, the gradients are highly reproducible, if stirring is adequate and hydrostatic equilibrium is maintained. I n any system operating by hydrostatic equilibration, allowance must be made for differences in density. This is most conveniently done by reducing the volume charged to each chamber in proportion to its density relative to the lowest density in the system. The error thus introduced is small for most cases involving aqueous buffers, but large deviations from the theoretical can be espected when liquids of widely different 860

ANALYTICAL CHEMISTRY

Nine sections of Lutite tubing ( 6 X 2'/4 I.D., '/a inch wall) cemented to top surface of plate B and to each other B. Lucite plale "8 inch thick. Nine-sided polygon inscribed in 10inch circle. Holes in central area for ventilation and for motor shaft not shown C. Holes (3/le-inch diameter) drilled halfway into plate B to connect with sides of hole for stopcock D (see insert) D. Eight Teflon stopcock plugs snugly fiiting 3/8-inch reamed holes. Each plug fltted with neoprene O-ring and Lucite handle. Hole ('/g-inch diameter) through plug connects vertical holes C (see insert) E. Recesses ( 1 '/*-inch diameter) extending '/2 inch from lower surface of plate B. Magnet assemblies, K, turn in these spaces F. Motor (KYC-23RB, 150 r.p.m., Bodine Electric Co., 2288 West Ohio St., Chicago 12, Ill.) and condenser mounted on Lucite block held to plate B by two long-shonked thumbscrews. Shaft fitted with coupling to engage I within 3/4-inch hole in center of plate B G. Base o f driving unit. Luciie ( 3 / 8 inch thick); same shape os plate B H. Central driving disk, 51/4 inches, '/4 inch thick 1. Mullilaw coupling (FA5 Boston Gear Works, Quincy 71, Moss.) to engage similar component attached to shaft of motor (see F ) J . Disk, 2'/4-inch diameter, attached to shaft. Brass washer between disk and bearing K. Lucite shaft ( 1 3/16 X l / d inch diameter) turning in bearing ('/c-inch I.D., 6/8-inch O.D., Fafnir S1 K7) mounted in base. Alnico V magnet ( 1 X '/8 inch, Arnold Engineering Co., Marengo, Ill.) inserted through hole a t top end of each shaft and held in place by section of soft plastic tubing compressed between disk and magnet L. Three Lucite spacers, '/2 inch high, cemented to G. Holes, 3/g-inch diameter, in top surfaces accommodate loosely three inch diameter) attached i o slightly tapered Lucite pegs (l/4 X bottom of plate B M. Cover, 10 X '/g inch Lucite disk with 43/a-inch diameter hole in center for ventilation, to resi loosely on chambers, A N. Lucite peg ('/c-inch diameter) attached to underside of cover, M, to tlt into inside angle between two cylinders when cover is in place 0. Lucite rod (1 X '/2 inch diameter) with central hole to position loosely tltting glass or plastic outlet tube at outermost point of inner wall of chamber

A.

densities are employed. Hon ever, this does not preclude the use of hydrostatic equilibration in such cases, since the resulting gradients are adjustable and reproducible If rigorous adherence to the theoretical curve is required, hydrostatic equilibration can be replaced with uniform piston displacement. This should be particularly advantageous for applications on a micro scale, \\-here hydrostatic equilibration can be expected to offer difficulties. K i t h or without such modifications, devices based on the principle described may find application in other fields. Precisely controlled density gradients of many shapes should be useful in ultracentrifugation studies of rapidly sedimenting particles. Also, the ability to produce an unlimited number of independent gradients of different molecular species suggests applications in investigations concerned with the determination of optimal conditions in multicomponent systems. DESIGN AND CONSTRUCTION

Devices based on the principle described can take many forms, but certain requirements must be met in the design and construction of each, if performance is to follow theory. The chambers should be as nearly identical as possible with respect to cross-sectional area a t any given level. If variation must be allowed, the cham-

bers should be arranged in a sequence in which the diameters increase in the direction of increasing density of liquid charged, so that the errors introduced by volume variation will reduce those resulting from the initial hydrostatic adj ustment . The channels connecting the chambers must be large enough to permit rapid hydrostatic equilibration and to prevent blocking by very small air bubbles, yet their volume should be insignificant relative to the volume of the adjacent ve9sels. The channels must be lower than the floor of the chambers in order to prevent the laminar exchange of liquids of differing density n-hich would otherwise occur even with the chamber contents in hydrostatic equilibrium. Mixing a t the bottom must be very efficient, and the height of the chambers should not be so great as to prevent adequate vertical transfer. Vortices should be kept as small as possible, consistent with adequate stirring, and all the vortices should be similar in size. Small stirrers made of soft iron wire-e.g., 1-inch sections of paper clip v-ire-covered n-ith thin glass tubing from melting point tubes are recommended for use in the chambers of magnetically driven mixers. Polyethylene-covered stirrers have a tendency to develop a rough surface after long use, with consequent occasional sticking. The distance between the stirrers and the driving magnets

inch wall A. Nine sections of Lucite tubing ( 3 l / 4 X 1 '/s I.D., cemented to plate 8 and to each other. Outer walls milled flat a t points of contact to reduce distance between inner walls of adjacent chambers to '/a inch B . Lucite plate, 6'" inch thick, 5'/~-inchdiameter C. Hole 1 3/16-inch diameter) drilled through plate B. Channel, 3/16 inch wide and l / 1 6 inch deep, milled into bottom surface to connect holes in adjacent chambers D. Lucite strips ( 1 l/g X 3/8 X '/I6 inch) cemented to underside of plate B to enclose channels connecting holes E. Three ' / 4 X '/z inch diameter Lucite pegs cemented to underside of plate B F. Aluminum disk ( 5 l / ~X '/g inch) attached to '/2-inch diameter steel rod. Pegs E fit loosely into 17/32-inch holes in disk G. Thumbscrews, with locknuts, to level mixer H. Haydon heavy duty inductor timing motor, 30 r.p.m. (Haydon Manufacturing Co., Torrington, Conn.) I. Lucite disk (5l/8 X ' / 4 inch) fitted with bearings ('/?-inch I.D., j/s-inch 0. D., Fafnir S l K 7 ) . Leveling bubble on upper surface and two '/a-inch pegs on lower surface (to position disk) are not shown J. Lucite driving disk, about 3-inch diameter and '/g inch thick, with straight knurl on edge. See text K. Section of rubber tubing forced onto shaft of paddle to serve as friction disk (about inch diameter). Separated from beoring b y brass washer 1. Lucite ~ a d d l e ( 3 ~ / 1X 6 5 / g X '/I6 inch, 45' twist) cemented to notched end of 1 X '/4 inch diameter Lucite shaft. Teflon washer between top o f paddle and underside of disk I M. Lucite rod (about 7/3z-inch diameter) to match, in the unstirred chamber, the displacement of liquid by paddles in the other chomberr N. Lucite rod ( 3 / 8 X 5 / 1 6 inch diameter) with central hale to accommodate glass or plastic outlet tube. For clarity, this i s shown in position above chamber 8; in normal use the tube would be inserted into chamber 1

I I I I I il

G-i3 Figure 7.

should be as sniall as possible to afford a strong, reliable coupling. Although mixing is not required in the last chamber, the liquid in this chamber should be stirred in the same way as that in the others, in order to avoid, in magnetic mixers, inequalities in contribution as the vortices increase with decreasing depth of liquid. This is of little importance in paddle-stirred mixers, but these require provision for tht. displacement of liquid in the unqtirred chamber to match the volume displaced by the paddles in each of the other chambers. Figure 6 illustrates the construction details of a nine-chambered Varigrad having a capacity of 3400 ml. Lucite was used, because this mixer n a s intended for use with aqueous solutions; other materials should be employed when organic solvents are involved. The central driving disk, H , is made of sponge rubber supported between two 5inch Lucite disks, the l o m r '/g inch thick and the upper 1/16 inch. This sandwich rests on a brass disk (2 X inch) soldered to a brass shaft having a '/Tinch diameter below the disk to fit a Fafnir S5K bearing 112 inch in inside diameter, and 11j8 inches in outside diameter and a 'I4-inch diameter above it to fit the coupling. Compression of the sponge rubber when the coupling is attached binds the disk assembly to the shaft. Very serviceable outer disks, J, have been made b r sandwiching three or four pieces of heavy canvas -between two %inch disks of Lucite. the lower of

(I

Paddle-stirred model

which (I/g inch thick) \vas cenicrited to the '/d-inch diameter Lucite ..haft. Foam or sponge rubber can also be used. Polyurethane foam wore down too rapidly. Presumably, plastic or fiber gears can be used instead of friction disks, if suitable sizes are available.

brass hub. The sniall outer disk< n c w sections of heavy-walled rubber tubing of suitable diameter. The pressure between the large and sniall disks should be just sufficient to assure contact a t all times. Excessive pressure puts an unnecessary load on the marginally powered motor.

I n the construction of the driving mechanism care should be taken to provide adequate space between the rotating disks and the l o w r surface of the plate mourited above them, but the magnets should be as close as possible to the tops of the recesses in the plate. It is advisable to drill a fern large holes through the central area of the plate to aid the circulation of air past the motor, A paddle-stirred model n i t h a capacity of 430 ml. is shown in Figure 7 . Lacking stopcocks, it is simpler to construct but less convenient to use.

The mixers shon-n in Figure3 6 and 7 were designed for use a t atmospheric pressure, with punips feeding the eniergent liquid to chromatographic columns. For this reason simple, loosely fitting covers \\-ere employed. If, however, flow iq to be induced by air presvre or by gravity with a llariotte-tube control, airtight covers fitted with connecting channels must be provided. I n preparation of the T'arigrad for us?, the channel' should be flushed with water to diiplace air before the stopcocks (lightly greased with silicone lubricant) are closed. After the chambers have been charged Iyith their respective solutions, the vertical connecting holes should be examined for air bubbles. It is advisabk, moreover, to watch the liquid levels for a short period after the start of an experiment until it beconies evident that the meniscus is falling in every chamber. K h e n a stopcock is being greased, it is necessary to open a very small vent hole, extending from the innermost tip of the stopcock hole to the bottom of plate B (Figure 6), to permit the escape of occluded air and excess grease. This vent must then be resealed with a drop of

While the chambers are being filled, the connecting channels are stoppered with eight 4 X 3/32 inch stainless steel rods fitted a t the lover end with a short section of 3 '16-inch diameter vinyl tubing n-hich projects about '8 inch beyond the steel. Pressure on the rod after insertion of the tubing into one of the holes forces the slightly tapered end of the rod downward, compressing the> tubing against the side of the hole. As in the case of the magnetically stirred mixer, the details of construction of the driving mechanism can be varied. I n existing models a Lucite disk inch thick has been used for J , supported on a 2-inch brass disk attached to a l/4-inch

VOL. 31, NO. 5, MAY 1959

861

plastic cement. If silicone grease is employed, regreasing will be Only infrequently required. Because the major function of the stopcocks is the isolation of the chambers during the short filling operation, slow leaks between the chamhers are of no consequence. The O-rings effectively prevent leakage to the exterior* On the Other hand, if One Or more of the chambers is to remain empty during an experiment, the closed stopcock should be supplemented with a stopper of the type described above for use in the absence of stopcocks.

LITERATURE CITED

(1) Alm, R. S., Williams, R. J. P., Tiselius, A., Acta Chem. S c a d 6 , 826 (1952). (2) Bock, R. M., Ling, N*-S.,ANAL* CHEM* 26, 1543 (1954). (3) B ~ H. G., ~ ~ i~ ~ ~et~h~ i i~, .~ ~ c t 16, a 245 (1955). (4) Busch, H., Hurlbert, R. B., Potter, V. R., J. Biol. Chem. 196, 717 (1952). (5) Gherkin, A., Martinez, F. E., D u m , hl. S., J . Am. Chem. SOC.75,1244 (1953). (6) Desreux, V., Rec. trav. chim. 6 8 , 789 (1949). (7) Donaldson, K. o., Tulane, v. J., Marshall, L. M., ANAL. CHEM. 24, 185 (1952). (8) Drake, B., Arkiv Kemi 8 , 1 (1955).

(9) Lakshmanan, T. K., Lieberman, S., Arch. Mitchell, Biochem. Biophys. K., Gordon, 45, 235 M.,(1953). Haskins, F. A., J . Bid. Chem. 180, 1071 (1949). ~(11)h Parr, ~ c. ~ W., . Proc. Biochem. SOC., 324th Meeting, xxvii (1954). ('?{&$' K* A'1 ANAL' 28, 1451

( l ~ ? a ~ ~ ~ ~ ~ ' , ~ ; l Fedl f i e ~ ~ ~ ~ RECEIVED for review November 19, 1958. Accepted February 9, 1959. Presented in art before the Federation of American Eocieties for Experimental Biology, Philadelphia, Pa., April 1958.

Automatic Derivative Potenti ometric and Spectrophotometric Titrations of Organic Acids H. V. MALMSTADT and D. A. VASSALLO' Deparfrnenf of Chemistry and Chemical Engineering, University of Illinois, Urbana, 111.

b Nonaqueous titrations of carboxylic acids, sulfonamides, imides, mercaptans, phenols, and enols can b e performed rapidly and accurately with a single titrant-solvent combination and automatic derivative potentiometric or spectrophotometric end point termination. The acids are titrated in acetone solvent with tri-n-butylmethylammonium hydroxide titrant. The applicability of many electrode pairs and indicators to potentiometric and spectrophotometric procedures, respectively, was investigated, and the most suitable ones applied to the automatic titration of a variety of acids. The end point reproducibilities are within 0.01 mi. of titrant, the buret reading error, and the results show that the automatic derivative procedure is an accurate method for determining percentage purity or neutralization equivalents of organic acids.

R

literature has contained several reports on the titration of acids in relatively inert solvents (1-4, 6, 7). Most titrations of organic acidic materials are currently performed by recording the titration curves ( I ) or by a manual potentiometric procedure (4). Investigations in this laboratory of the automation of titrations by a derivative technique (10-15) demonstrate that a wide variety of acids can be titrated rapidly and accurately using acetone solvent, tri - n butylECENT

-

Present address, Polycheniicals Department, Experimental Station, E. I. du Pont de Nemours & Co., Inc., Wilmington, Del. 862

ANALYTICAL CHEMISTRY

methylammonium hydroxide titrant, and either automatic derivative potentiometric (10) or automatic derivative spectrophotometric ( I S ) end point termination. Many electrode systems were studied to determine which was most suited for automatic derivative potentiometric end point detection in acetone solvent. A platinum (10% rhodium) indicator electrode us. graphite reference electrode connected directly to the input of the commercially available SargentMalmstadt automatic titrator (15) was suitable in titrating automatically all except the very weakest acidic materials, such as phenol. Such materials can be titrated automatically using azo Violet indicator and the derivative spectrophotometric end point detection method. Other indicators were studied in acetone solvent and the most suitable were selected and tested using automatic derivative spectrophotometric end point termination. The same indicators are suitable for manual visual end point detection in acetone solvent. The precision of the automatically obtained results was generally rrithin 0.01 ml. of titrant, the reading error for the buret used. The automatic titration procedure is rapid and accurate for determining percentage purity or neutralization equivalents for organic acids. Fritz and Yamamura (4) used acetone and Bruss and Wyld ( I ) used methyl isobutyl ketone (4-methyl-2-pentanone) as solvent for the titration of organic acidic materials using quaternary ammonium hydroxides as titrants. hlalmstadt and Vassallo described the use

of acetone-water solvent for the quantitative detennination of perchloric-acetic acid mixtures ( I S ) and sulfuric-hydrochloric and sulfuric-nitric acid mixtures (19). Acetone was selected for this study because of its availability, low cost, purity, and general solvent characteristics. APPARATUS AND MATERIALS

TITRATOR. Automatic derivative potentiometric and spectrophotometric titrations can be performed using a modified titration stand ( I S ) for the Sargent-Malmstadt titrator, but the new commercially available SpectroElectro titration stand (E. H. Sargent &- Co., Chicago, Ill.) is more convenient for switching to either potentiometric or spectrophotometric end point termination, with simple polarity, electrode and filter selection, and an easy method for the insertion and removal of the titration vessel. The photoconductive detector circuit was used for the spectrophotometric procedure and the platinum (10% rhodium) us. graphite electrode pair for the potentiometric procedure. RECORDER. Model G-10 Varian (Varian Associates), 100-mv. recording potentiometer. BURET. A 10-ml., self-zeroing, Teflon stopcock buret (Fischer & Porter Co.) was filled with titrant from an elevated bottle through the side arm which was fitted with a Teflon stopcock, and all interconnections were made with either Teflon or polyethylene tubing. BURETVALVE. Gum rubber tubing with a pinch-off solenoid is commonly used for starting and stopping the delivery of titrant, but the tubing is attacked by the nonaqueous titrant and requires frequent replacement.