Determination of the Drying Rates of Thin Films c.
GEORGE RIEGER~AND s. GROVE, J R . ~ University of Minnesota, Minneapolis, Minn.
An apparatus, which can easily be constructed from equipment available in any laboratory, has been devised for measuring the drying rates of thin films. This apparatus and technique have been applied to a study of thin nitrocellulose films, the maximum dry weight of which was about 0.02 gram; the area was about 13 sq. cm. (2 square inches) ; the total solvent content was about 300 mg.; and the drying time was about 20 minutes. The drying rates of these selected films have been calculated and the results presented in graphs. The suggested apparatus and technique have been shown to give satisfactory and reproducible results.
ITu’ of
T H E past few years there has been a revival of interest in the evaporation rates of solvents and in the determination these evaporation rates. Most of this work has been done on pure solvents, and there have been very nearly as many different types of apparatus employed as there have been investigators. Each new development has been in a sense a refinement of older methods in an attempt to obtain an easier operation and a greater reproducibility of results (1, 6, 6, 8, 10, 11, I d ) . From an over-all viewpoint it has been the endeavor of this investigation to devise an apparatus that will minimize the effects of any local operating conditions and give consistent results. Although these results may not quantitatively be comparable to results obtained on other apparatus, a t least they will be relative. In the case of a pure solvent, the more accurate evaporation measurements show that a plot of the per cent solvent evapo1 2
Present address, Ethyl Gasoline Corporation, Baton Rouge, La. Present address, E. I. du Pont de Nemours & Company, Buffalo, N. Y.
rated us. time is a straight line; this means that a pure solvent has a constant evaporation rate. If one or more solvents are introduced into the original solvent and evaporation measurements are made, then the plot of the per cent volatiles evaporated is no longer a straight line. This means that the solution containing two or more solvents does not have a constant evaporation rate. If instead of another solvent or other solvents, a nonvolatile component be added to the original solvent, it has been shown that the resulting solution again exhibits no constancy of evaporation or drying rate. Bogin and Wampner (S), Stewart and Dorsch (9), and Durrans and Davidson (4) have done much work on the drying rates of solutions of nonvolatiles in volatile solvents. Their published reports indicate that the presence of nonvolatile materials affects the drying rate of a solution and that this rate falls off as the solvent evaporates, finally approaching zero as the resulting mass nears dryness. The determination of the drying rates of such solutions means the employment of some apparatus that will enable weight readings to be taken a t various intervals of time during the drying process. Durrans and Davidson placed a quantity of their solution (nitrocellulose in a suitable solvent) in a shallow dish and passed air of a constant temperature over it. The sample was large enough and the amount evaporated in each time interval of sufficient magnitude, so that any evaporation occurring during the time of weighing probably introduced very little error. However, the use of a large sample of any appreciable depth means that an actual thin film as ordinarily dealt with in the application of a protective coating is not simulated. Information on the drying rates of any sort of film-forming solutions, it seems, should be obtained when the solution is in the form of a thin film. The determination of the drying rate of a thin film, however, brings forth difficulties not ordinarily encountered when larger and deeper volumes of solutions are considered. For example, the supporting base or tare on which the film is cast generally constitutes the major portion of the total weight of the whole system. Therefore, a very sensitive weighing de-
FIGURE1. OPTICALSYSTEM D. Pointer Balance lens F. Projection screen
A. Light source E. Condenser lene C. Pointer scale
E.
326
April 15. 1942
ANALYTICAL EDITION
321
J
11
31111 E:
I
FIGURE 2. WIRINGDIAGRAM A. B. C. D. E.
vice must be used t o detect small changes in the weight of the film.
Transiormer
Push-button switch Single- ole double-throw switah 50-CD. kkment 3 2 - 0 ~filament .
which held in place a length of rubber tubing connected to an aspirator. Thus the air in the balance was gently exhausted at
Apparatus The requirement of sensitivity suggested the use of some standard analytical balance. However, using a chainomatic type of analytical balance, it was found that with very thin films which weighed but a few milligrams and contained fast-evaporating solvents, the balance operator could not keep up with the weight changes with any degree of accuracy. In an attempt to overcome this operational difficulty, the following apparatus w s devised for determining the drying rates of thin films. A Christian Becker (Type 508-A) air-damped analytical balance was used. The distinguishing feature about this balance, which makes it or one similar to it particularly adaptable to a study of the drying rates of thin films, is the scale attached to the pointer. This scale, similar to a microscopic stage micrometer, is ordinarily read by means of a low-power microscope attached to the body of the balance caae. It is graduated in milligrams up to 100 mg. on each side of the zero mark. Each milligram is divided into two parts, so that it is possible to estnnate to 0.1 mg. weighing with the same accuracy that a rider or chain o m be read on the usual type of analytical balance. The beam is air-damped, which under normal operating oonditions prevents overshooting and provides a smooth response to weight changes in the pans. It is possible to read this balance over the range of a ZOO-mg. change in weight without changing the weights on the pan. The hairline used in conjunction with the pointer scale is an integral part of the eyepiece of the microscope attachment. Since, in general, it is fatiguing to peer constantly through such a microscope the objective lens was adjusted so that the scale could be prdjc screen. In or< the pointer s necessary to 1 with a small beam of light source empl double-filamen headlamp bulb and a 50-cand ment. The50 filament was oneration, wl candlepow& i kept wired fol of failure of and wiring c shown inFiglu The sereen 7 on-one edge that supporte( and wsSSo con it could be swu: of the balance ' were being ch the case. Th pane of the bal removable. 'I raised sliebtlv neath it