1NTBRPRETAHON or ULTRASTRUCTÜRB
135
was the morphology seen most of the time. Without entering into a detailed discussion of
the reasoning (see Rebhun 1965) we consider this to be due to a crystallization phenomenon
of some sort. The second type of morphology seen is one which in many respects resembles
that of chemically xed material and is illustrated, for example, in Figures 4, 5, etc. The
problem which faced us was that in the majority of cases, the morphology obtained after
freezing was that of ice crystal artefact, whereas, only in a small, and variable number of
cases was the other morphology obtained. We wished to investigate the sources of this
variability- There were three things which seemed obvious to us to try. The easiest was to
reduce the size of the tissue pieces (or egg sample size in the case of marine eggs) so as to
reduce heat diffusion paths and obtain more rapid freezing. We ultimately froze tissue
pieces about 0.1 to 0.2 mm maximum diameter and egg samples of at most 0.01 ml spread
in a layer one egg thick (about 60 microns). The next technique investigated was the method
of rapid freezing and the quenching uid used, the attempt again being to obtain the fastest
freezing possible. The nal variable, was the temperature of dehydration. This was important since, even though we might freeze tissue in an ice crystal free form, the possibility
existed that at the usual dry—ice temperatures used for dehydration, devitrication, that is,
Spontaneous transition from the glassy state to the crystalline state, might occur. For pure
water this occurs at about —— 150 to — 160 °C. (Dowell and Rinfret 1960). Even though one
would not expect this low temperature to hold for tissue, since devitrication points for
various solutions are usually much above this temperature (Luyet, et al, 1958), neverthe—
less, in the absence of data for tissue, the lowest practical temperatures were used. Devi—
trication and subsequent recrystalhzationwere shown by Meryman (1957) to result in
large ice crystals in model systems consisting of protein solutions.
We will now present the work on transient cooling in some detail.
A. Transient Cooling Investigations
In 1961, Cowley, Timson and Sawdye published a paper in Biodynamica, which showed
that very considerable increases in cooling rate could be obtained in liquid nitrogen if
insulator coats or layers of particles were properly applied to the surfaces of objects to be
cooled. This is explicable in terms of what is known of heat transfer to boiling uids (see,
e.g., Westwater, 1959 or Rohsenow and Choi, 1961). Essentially,.particles on the surface
act as nucleating sites for bubble formation so that instead of getting a continous lm of
gas surrounding the object if the object temperature is much above the maximum nucleate
boiling point of the uid, one obtains a myriad of bubbles each of which carries away latent
heat of vaporization, as it were, and also creates turbulence, both effects resulting in increas—
ed heat transfer. The effect of insulator coats is more complex but is not too diîcult to
understand. We began a systematic search for the best quenching uids using all the reason—
able hydrocarbons and uorcarbons we could nd. Specically, we wished to compare
freezing rates in uids, such as isopentane, whose boiling point is about room temperature .
and therefore would not show an increase in freezing rate with either insulator coating Or
powder, those whose boiling point is not far from 0 -°C, where the coating technique might
be of marginal use, and those whose boiling points were far below 0 °C where insulator
coats and powders could be expected to be of considerable use. We hoped to use microthermocouples embedded in tissue such as Stephenson
had used. It soon became
clear, however, that constructing such thermocouples (to have measuring elements in the
*neighborhood of 2.5 microns), was quite diicult to do reproducibly andwould not allow
135
was the morphology seen most of the time. Without entering into a detailed discussion of
the reasoning (see Rebhun 1965) we consider this to be due to a crystallization phenomenon
of some sort. The second type of morphology seen is one which in many respects resembles
that of chemically xed material and is illustrated, for example, in Figures 4, 5, etc. The
problem which faced us was that in the majority of cases, the morphology obtained after
freezing was that of ice crystal artefact, whereas, only in a small, and variable number of
cases was the other morphology obtained. We wished to investigate the sources of this
variability- There were three things which seemed obvious to us to try. The easiest was to
reduce the size of the tissue pieces (or egg sample size in the case of marine eggs) so as to
reduce heat diffusion paths and obtain more rapid freezing. We ultimately froze tissue
pieces about 0.1 to 0.2 mm maximum diameter and egg samples of at most 0.01 ml spread
in a layer one egg thick (about 60 microns). The next technique investigated was the method
of rapid freezing and the quenching uid used, the attempt again being to obtain the fastest
freezing possible. The nal variable, was the temperature of dehydration. This was important since, even though we might freeze tissue in an ice crystal free form, the possibility
existed that at the usual dry—ice temperatures used for dehydration, devitrication, that is,
Spontaneous transition from the glassy state to the crystalline state, might occur. For pure
water this occurs at about —— 150 to — 160 °C. (Dowell and Rinfret 1960). Even though one
would not expect this low temperature to hold for tissue, since devitrication points for
various solutions are usually much above this temperature (Luyet, et al, 1958), neverthe—
less, in the absence of data for tissue, the lowest practical temperatures were used. Devi—
trication and subsequent recrystalhzationwere shown by Meryman (1957) to result in
large ice crystals in model systems consisting of protein solutions.
We will now present the work on transient cooling in some detail.
A. Transient Cooling Investigations
In 1961, Cowley, Timson and Sawdye published a paper in Biodynamica, which showed
that very considerable increases in cooling rate could be obtained in liquid nitrogen if
insulator coats or layers of particles were properly applied to the surfaces of objects to be
cooled. This is explicable in terms of what is known of heat transfer to boiling uids (see,
e.g., Westwater, 1959 or Rohsenow and Choi, 1961). Essentially,.particles on the surface
act as nucleating sites for bubble formation so that instead of getting a continous lm of
gas surrounding the object if the object temperature is much above the maximum nucleate
boiling point of the uid, one obtains a myriad of bubbles each of which carries away latent
heat of vaporization, as it were, and also creates turbulence, both effects resulting in increas—
ed heat transfer. The effect of insulator coats is more complex but is not too diîcult to
understand. We began a systematic search for the best quenching uids using all the reason—
able hydrocarbons and uorcarbons we could nd. Specically, we wished to compare
freezing rates in uids, such as isopentane, whose boiling point is about room temperature .
and therefore would not show an increase in freezing rate with either insulator coating Or
powder, those whose boiling point is not far from 0 -°C, where the coating technique might
be of marginal use, and those whose boiling points were far below 0 °C where insulator
coats and powders could be expected to be of considerable use. We hoped to use microthermocouples embedded in tissue such as Stephenson
had used. It soon became
clear, however, that constructing such thermocouples (to have measuring elements in the
*neighborhood of 2.5 microns), was quite diicult to do reproducibly andwould not allow