Using a modification of the vibroscope technique, studies have been made of the sorption kinetics of water vapor in single wool fibers. Small changes in water content, described as interval sorption and desorption, occur in two stages, the first occupying a few minutes and the second many hours (at room temperature). This behavior is similar to that recently reported for interval sorption of organic vapors in “glassy” polymers (e.g., acetone in cellulose acetate). The first stage of sorption is considered to be due to diffusion in accordance with Fick's law; the slow increase in concentration during the second stage is thought to result from partial relaxation of the swelling stress. Large changes in water content are described as integral sorption and desorption. In integral sorption the slow second stage of sorption is not observed; the explanation is thought to be that relaxation of swelling stress, and hence the second stage, are greatly accelerated by the transient stress associated with the steep concentration gradient present during integral sorption. The slow second stage is present in integral desorption, the probable reason being that in this case there is no steep concentration gradient. Integral sorption to a given value of relative humidity (R.H.) results in a higher equilibrium water content than if the same R.H. is reached in a series of interval steps. This effect is thought to be related to the hysteresis which is observed in the isotherm in a sorption‐desorption cycle; the transient stress present in integral sorption leaves the material in the same condition as if it had been brought down the desorption curve of the isotherm. The sorption hysteresis is thought to result from metastability of the interchain bonds which undergo relaxation in the second stage of interval sorption and desorption. All the effects observed are likely to be qualitatively characteristic of vapor sorption in any polymer at temperatures below the second‐order transition.

A number of interval sorption experiments have been made and analyzed, assuming Fick's law; diffusion coefficients were derived. These coefficient have been used to predict the behavior in integral sorption; the disagreement between prediction and experiment is discussed. Integral sorption experiments have also been carried out on horse hairs of various diameters to trace the influence of diameter on the sorption rate. GeneralThe general shape of the'water sorption curve by wcx~l for small changes of relative humidity is well known I 1], 22 ~ ; it is very similar to those for water in cethnoso 18, 20) and for organic vapors in cellulose acetate 2 ~ . These curves may best he doscrihed in terms of a two-stage sorption process in which the first stage corresponds to H irkian diffusion and the second stage to a further uptake of sorhate the rate of which is not controlled ltv diffusion hut hy a molecular relaxation process of the sorbent.The exact nature of this relaxation process is not yet fully understood, although an approximate calcutation 121 has shown that it may he due to the relaxation of the elastic strain imposed on the polymer network 1>v the volume swotting accompanying the firat stage of sorption.In an attempt to analyze the observed kinetics of water sorption by wool, perhaps the most outstanding feature which requires explanation is the difference in the form of the uptake curves for small and large steps in relative hunndity. In the absence of a more appropriate theory, an analysis in terms of Fick's law has to 1>e undertaken. It has been shown by Crank 151 that, at least yualitativelv, the major characteristics of the sorption behavior of many polymer systems can be explained in terms of 1: ick's s law provided the diffusion coefficient is made a function of (a) the concentration, (b) I the time. and (c) the strain set up in the sorbent.The only available data for diffusion coemcients which may be applicable to the wool-water systemhave been reported hy King 17 J for horn keratin using a steady-state permeability method. King found that the steady-state diffusion coefbcient was a function of the concentration of the water in the keratin and published the curve relating regain and diffusion roefhcient. L'nfortunately, the physical form of the wool fiber makes it impossibte to devise a steady-state experiment hy which the concentration dependence of the diffusion coefhrient can he measured. One is forced, therefore, to aplroximate the required conditions of the measurement hy making the step of concentration as small as possible and assuming that. over the small range used, the coefficient is constant This type of experiment, however, does not rule out time or ,train effects.The small-step sorption experiments reported here were undertaken with a twofold purpose in view: to check whether agreement could be ohtained between diffusion coefficients for wool keratin and Iving'~ published figures for horn keratin and to find to what extent the sorption and desorption curves predicted by the concentration-...

1NTRODUCTIONThe paper describes measurements made of the dynamic rigidity modulus of wool fibers during the period when the fibers are sorbing or desorbing water (from the vapor phase). Anomalous behavior is observed including reduced values of rigidity modulus. The behavior appears to provide direct evidence of stresses being induced during the sorption process. EXPERlMENTS AND DISCUSSlON OF RESULTSThe dynamic rigidity modulus of wool fibers was measured by the well-known method (see, e.g., Speakman') of using the fiber as the suspension of a torsion pendulum. The modulus n is related to the period of oscillation T , the length 1 and radius R of the fiber, and the moment of inertia I of the bob (all of which quantities can be determined) by the expressionThe method is well establishedlJ for determination of the equilibrium value of rigidity modulus of fibers under given conditions of regain, temperature, etc. I n the present series of experiments cleaned Merino wool fibers were used, of mean diameter 25 p and length 3.50 cm. The bob was a brass disk of mass 0.68 g. and moment of inertia 0.059 g. cm.2 rotating in its own (horizontal) plane; the weight of the bob was just sufficient to remove the fiber crimp. The period of the oscillations observed was in the range 17 to 33 sec.If the period of oscillation is measured while the suspension fiber is sorbing or desorbing water vapor, the effect of the sorption process on the rigidity modulus can be observed. Curve I of Figure 1 shows the result of such an experiment, in which a fiber which had been dried (0% relative humidity (R. H.) at 20°C. for 20 hr.) was suddenly introduced into an atmosphere of 61% R. H. For convenience relative rigidity rather than rigidity modulus has been plotted. The term rigidity is here taken to mean the restoring torque per unit angular displacement of the bob. Denoting this by N we haveRelative rigidity is defined where N o is the rigidity of the dry fiber and To the period of oscillation with the dry fiber as suspension. As abscissa (time"*) is used, chiefly for convenience in presenting the time range involved. The striking feature of the curve (Curve I) is the pronounced ('undershoot" followed by a gradual recovery toward the equilibrium value, The latter value agrees closely with published data. Figure 1 also shows the regain vs.curve (curve 11) obtained simultaneously during the experiment by the automatic vibroscope3 method.The pronounced dip in the curve of relative rigidity requires explanation. If relative rigidity is calculated for a series of values of regain, using eq.(2) and published equilibrium values for rigidity modulus' and fiber r a d i u~,~ no dip is observed. This curve is plotted in Figure 1 as Curve 111.(The slight rise in N / N o at the beginning is due to increasing diameter outweighing the falling rigidity modulus in this regain range.) It is not strictly correct to take equilibrium values of swelling vs. regain to calculate rigidity during the sorption process since the water is then not uniformly distributed ...

The increase in electrical conductivity of keratin fibers after the application of an abrupt change of the relative humidity from 0 → 90% RH has been examined as a function of time. A comparison of these results with regain vs. time, for the same experimental conditions, indicates that the water inside a fiber is not initially in a state which facilitates conduction. It is proposed that the slow rise of conductivity is due to the formation of a hydrogen bonded network which allows the passage of protons under the influence of the applied electric field.