A Versatile Method of Ambient-Temperature Solvent Removal

Abstract

Isolation of heat-sensitive reaction products in post-synthesis workup procedures often requires ambient-or low-temperature solvent removal. In the method demonstrated here, solvent evaporation is driven by the pressure gradient between a distillation flask and a chilled receiver in an evacuated closed system containing a minimal amount of residual noncondensable gas. Using an all-glass apparatus, the method is exemplified by evaporation of solvent samples from a distillation flask containing 50 mL of either dimethylformamide, dimethyl sulfoxide (DMSO), or N-methylpyrrolidone (NMP). The distillation flask is suspended in a water bath at temperatures of 18−28 °C, the evaporated solvent is collected in a receiver chilled with liquid nitrogen, and the entire process is completed in 90−140 min. The practicality of this method is further illustrated on a bench-chemistry scale by DMSO and NMP solvent removal from solutions of benzophenone, monitored by gravimetric and ¹H NMR methods. Modification of the demonstrated method to mimic freeze-drying conditions (by reducing heat flow to the distillation flask) can be used for recovery of water-soluble compounds including polymers and biopolymers. We propose the name “cryovap” for this solvent removal method.

INTRODUCTION

Workup procedures of many solution-based synthesis protocols commonly end with a solvent removal step. Aqueous solution-phase chemistry generates the least solvent waste, and unless the reaction product is oxygen-sensitive, water can be evaporated into the air or absorbed by a suitable desiccant. Other solvents are typically removed by evaporation followed by collection of the condensed solvent in a receiver. Relatively small amounts of volatile solvents with vapor pressures higher than the lowest pressure that the vacuum pump is capable of providing can be evaporated at ambient temperature, with vapors condensed in a solvent trap chilled with an appropriate cryogen.^1–3 For larger volumes, rotary evaporators are commonly preferred; BÜCHI guidelines suggest the water bath temperature to be set to 50 °C and the water cooling the condenser to be at 10 °C, while pressures are set for specific solvents, such that their boiling points are lowered to 30 °C (Δ20 °C rule).^4,5 In practice, few bench chemists are comfortable with removing dimethylformamide (DMF; B.P. 30 °C at P = 4.5 torr) using a rotovap because it is commonly paired with an oil-free diaphragm pump, and such a low pressure is difficult to achieve. Alternatively, raising the temperature of the heating bath usually requires using a heat-transfer liquid other than water, limits the lifetime of the seals used in the rotovap, and increases the possibility of product decomposition.

In 2000, Cherian reported the concept of high-boiling solvent removal via vacuum drying as a method developed for drug formulations.⁶ As the inventor states, “In this procedure, a high boiling solvent is removed from a solution of a pharmaceutical compound by adding a low boiling co-solvent and applying vacuum at a temperature greater than the freezing point, but lower than the boiling point of the solvent mixture. While the method of removal of the high boiling solvent is not entirely understood, it is believed that the low boiling co-solvent augments the mass transfer rate of the high boiling solvent.”⁶ The concept of vacuum evaporation under centrifugal force was realized in the Vaportec V-10 evaporator.⁷ The evaporator uses a high-speed (∼6000 rpm) motor to spin a vial containing a sample, creating a thin film of the solvent that can be readily evaporated from the heated vial, while the consequent centrifugal force prevents “bumping.” Further development on this system and inclusion of an additional external vacuum pump were attained in the Biotage V-10 Touch, which allowed removal of higher-boiling solvents such as dimethyl sulfoxide (DMSO) and N-methyl-2-pyrrolidone (NMP).⁸ A similar principle is utilized in Genevac centrifugal evaporators.⁹ High-vacuum distillation, or molecular distillation, is a technique for purification of substances with low volatility, having vapor pressures of 10⁻³ torr or less on a relatively small scale.³ The apparatus is designed so that the distance from the surface of the evaporating fluid to the condenser is less than (or comparable to) the mean free path of a molecule of the distillate vapor. It is imperative for molecular distillation that the residual non-condensable gas pressure is maintained at very low values, which is achieved by means of a high-vacuum (diffusion or turbomolecular) pump coupled with the freeze−pump−thaw technique.¹⁰ In some cases, the solvent from a reaction can be removed by extracting it with another solvent, provided that the targeted solute is insoluble in the extracting solvent.¹¹

Freeze-drying is a solvent evaporation technique that takes place at a temperature lower than the freezing point of the solvent. In this method, a frozen solvent sublimes under dynamic vacuum conditions and recondenses in the condenser chamber, thus providing a method for isolation of the solutes with no exposure to heat and with reduced residual solvent content. This technology employs freeze-drying machines, or so-called lyophilizers, paired with vacuum pumps.^12,13

In summary, there is a great variety of solvent evaporation methods relying on specialized equipment for solving specific technological issues. In particular, removal of high boiling point solvents remains a challenge requiring high-vacuum techniques and advanced equipment and is associated with higher costs. There is no universal technology that is equally efficient in different solvent systems, and hence, a need for further improvement exists.

In this paper, we describe a simple method of solvent removal in an evacuated closed system at ambient or, in some cases, low temperature. The method is suitable for solvents with a broad range of volatility, having normal boiling points of up to at least 202 °C (higher boiling point solvents are possible to remove with longer experimental times), and is only limited from a volume standpoint by the size of the glassware used. Furthermore, solutions containing volatile corrosive components (such as hydrogen halide) can be efficiently evaporated with minimal risk of damaging the hardware because this method is realized in an all-glass apparatus without using a rotovap or any of its components or a vacuum pump beyond an initial relatively short evacuation stage. This method is recommended for general solution-phase synthesis routines in organic, organometallic, biological, and coordination chemistry on a basic or applied research and development scale.

RESULTS AND DISCUSSION

Aside from temperature, surface area, and concentration, the rate of solvent evaporation from a solution depends on the total pressure of the atmosphere above it. One of the most well-known examples of this is the comparison of a desiccator to a vacuum desiccator: at equal temperatures and using the same amount of identical desiccants in each, water evaporation rates are very different between the two. The solvent evaporation method described here is based on this same principle. The best condition for efficient solvent evaporation can be achieved if the system is evacuated to a point where any noncondensable (at a temperature in which the condenser will be cooled) gas (e.g., air) is removed completely during the initial pumping stage. This can be accomplished if the pressure provided by the pump is lower than the vapor pressure of the solvent and if pumping is performed long enough so that the solvent vapor expels the noncondensable gas. Alternatively, if the solvent vapor pressure is lower than the pressure provided by the pump, a minute amount of inert cosolvent with a higher vapor pressure can be added. In this case, most of the cosolvent evaporates during the initial pumping stage, expelling the remaining noncondensable gas. At this point, the vacuum stopcock is closed (Figure 1), and the system is isolated from the vacuum manifold (not shown). Once the condenser/receiver is chilled, the remaining traces of the vapor of the assisting solvent condense, allowing the magnitude of the internal vacuum to reach the vapor pressure of the main solvent. Since the solvent vapor pressure in the chilled receiver is lower than that in the distillation flask, a condition of mass transfer is reached and distillation takes place, provided that the surrounding medium (air or water bath) supplies thermal energy to the evaporating solution.

Figure 1.

Ambient-temperature vacuum distillation system.

The apparatus can be assembled in a variety of different configurations. The terminology used for describing its various components is analogous to that used for common distillation setups. A Schlenk flask is a convenient choice as a distillation flask, whether or not the solute is air-sensitive, because it eliminates the necessity to use a separate stopcock as an evacuation port. Two types of condensers/receivers are shown in Figures 2 and 3. For simplicity, they will be referred as “receivers”. The receiver in Figure 2^a is not commercially available, so it must be ordered from a custom glassblowing shop, but the alternative type in Figure 3 is an assembly of a widely available distillation bridge and a round-bottom flask (a bottom-sealed tube with a ground-glass joint may be used in place of the flask).

Figure 2.

Distillation apparatus with a cylindrical condenser/receiver. The receiver shown is filled with a frozen solvent.

Figure 3.

Distillation apparatus with a bridge-connected condenser/receiver.

In order to evaluate the performance (distillation rate) of the described method, four solvents were selected (their normal boiling points are in parentheses): water (100 °C), DMF (153 °C), DMSO (189 °C), and N-methylpyrrolidone, NMP (202 °C). The three non-aqueous solvents, DMF, DMSO, and NMP (50 mL of each), were distilled from a 200 mL Schlenk flask, temperature-equilibrated with a water bath at 18−28 °C, into a receiver chilled with liquid nitrogen (Figure 4, Examples 1–3). The distillation time ranged from 90 to 140 min. Water (50 mL) was also distilled from a 200 mL Schlenk flask, temperature-equilibrated with a water bath (21−22 °C) (Figure 5, Example 4). In this case, the receiver was chilled with ice, and it took 3 h to complete the distillation.

Figure 4.

Vacuum distillation with a liquid nitrogen-chilled receiver.

Figure 5.

Vacuum distillation with an ice-chilled receiver.

The important methodology considerations are as follows. The main principle of the described method is that the pressure in each part of the apparatus must be equal to the vapor pressure of the solvent being distilled. In a typical apparatus (Figures 1−3 and 7), the evacuation port is located between the distillation flask and the receiver. If during setup the flask is charged with a solution but the receiver is empty, evacuation would cause the noncondensable gas to be expelled by the solvent vapor only from the distillation flask but not the receiver (this is the case in a typical evacuation procedure). It is imperative that a minute amount of the same or assisting cosolvent be added to the receiver before evacuation so that the noncondensable gas is removed from both sides of the apparatus. This is the step that allows for complete removal of the noncondensable gas from the system, a condition not met in traditional solvent removal techniques. The amount of cosolvent (if used) can be estimated from its vapor pressure and the total volume of the apparatus; the pumping is performed until the cosolvent evaporates fully from the receiver.

Figure 7.

Vacuum freeze-drying distillation.

Absolutely essential to the success of this method is a leak-proof assembly of the apparatus. To ensure that this condition is met, all of the ground-glass joints and stopcock must be of the highest quality. Individually ground stopcocks (with a numerically matching stopcock body and plug) and Schlenk flasks with such stopcocks are available from many glass companies, and each pair of inner/outer joints should be carefully matched or individually ground. Similarly, these joints must all be greased with great care using high-vacuum grease (silicone-or hydrocarbon-based) and making sure that each joint and stopcock are completely transparent; a rotation test of each greased joint should be smooth and show no smeared air bubbles.

Rapid stirring of the solution being evaporated is another important aspect. The main benefit of high-rpm stirring is that it helps prevent violent boiling of the solvent and consequent bumping, which is especially common during the initial evacuation step. Additionally, rapid stirring causes the surface area of the evaporating liquid to increase, which in turn speeds up evaporation.

As a cryogen, liquid nitrogen is the most efficient and the one to be used for solvents with a low vapor pressure. Most solvents will freeze in the receiver at the temperature of liquid nitrogen (Figure 2), which is not a problem if a relatively small amount of solvent is to be removed. When working with larger volumes, the solvent tends to freeze in the upper part of the receiver, plugging it and hampering further evaporation. This problem can be addressed by having a Dewar flask charged with a cryogen to one-fourth to one-third of its capacity so that only the lower part of the receiver is at the temperature of the cryogen. The upper part of the receiver, which is still inside of the Dewar flask, also contributes to vapor condensation but is less likely to accumulate a frozen solvent and become clogged. As the distillation progresses, more cryogen can be added or the Dewar flask can be repositioned higher so as to submerse more surface area of the receiver in the cryogen (Figure 4). A good indication that it is time to do so is when the solvent begins to condense in the upper part of the receiver. If the distillation is left unattended for an extended period of time and then found unfinished, then the cryogen (liquid nitrogen) should not be refilled because the liquid solvent that is likely to be present in the receiver in bulk amount will freeze, possibly causing the receiver to burst. This is especially the case if the frozen solvent has a lower density than that of its liquid counterpart, as in the equilibrium between water and ice.

For evaporation of large volumes of relatively volatile solvents, it may be beneficial to select cryogens so as to prevent freezing of the distillate. Otherwise, the vapors condensing and quickly freezing primarily in the top part of the receiver (near the cryogen’s upper level) often make a clog, and the lower part of the receiver remains empty through the rest of the distillation process. For water (or aqueous acid) removal, ice appears to be the cryogen of choice (Figure 5, Example 4) because it is not cold enough to freeze the distilled solvent.

In order to further illustrate the practicality of the described method, we examined solvent evaporation and solute recovery from solutions of benzophenone in DMSO and, separately, NMP. The change in initial solute mass was used as a control for the recovery, and ¹H NMR spectrometry was used for quality control. Benzophenone was used as a solute because of its high solubility in both DMSO and NMP and in CDCl₃ for NMR sampling (and because its ¹H NMR chemical shifts are distinctly different from the chemical shifts of these solvents). It was also chosen due to its stability, availability in pure form, and relatively low volatility (normal boiling point of 305 °C). Our results showed different behaviors between solutions in DMSO (Example 5) and NMP (Example 6). The chilling effect of rapidly evaporating DMSO from a dilute benzophenone solution results in its partial freezing. However, as the concentration of the solute rises, the freezing point of the solution drops, causing the sample to liquefy. This likely means that the melting diagram of benzophenone−DMSO mixtures is of a eutectic type. As expected, in each trial, after the samples initially froze and thawed, they remained liquid through the rest of the experiment, and the residue collected and tested by ¹H NMR after DMSO removal proved to be pure benzophenone with no visible traces of DMSO (Figure 6). The water bath temperature fluctuated between 19 °C initially and 26 °C at the end over the course of the 80−85 min evaporation time, and the resulting mass loss of benzophenone was in the range of 9−12%, depending on the specific combination of temperature and time in different trials. In reality, the temperature of the sample is always lower than that of the water bath due to the endothermicity of evaporation, and this is evident from the observation that DMSO (M.P. 17.9 °C) freezes at a water bath temperature of 23−24 °C even with high-rpm stirring to promote heat transfer. The recovered benzophenone samples were in a supercooled liquid state, which solidified at room temperature after an unpredictable amount of time. After solidifying, the samples’ measured M.P. was 47−47.5 °C (lit. M.P. 47.9 °C).

Figure 6.

¹H NMR spectra of benzophenone recovered from solutions in DMSO (top) and NMP (bottom). Only the bottom spectrum is integrated because it contains leftovers of the solvent. Both spectra were taken in CDCl₃.

As in DMSO, benzophenone’s solubility in NMP is very high, and due to the low melting point of NMP (−23.1 °C), these solutions remained homogeneous throughout most of the experiment in each trial until the NMP was almost completely removed. Unlike the DMSO solutions, the residue from the solutions in NMP solidified at the end of solvent evaporation, which hampered residual solvent removal as evidenced by ¹H NMR (Figure 6). After an evaporation time of 80−85 min in a water bath fluctuating between 24 and 26 °C, the residual NMP content (by NMR peak integration) was 4−10 mol % (2.2−5.7 wt %). The likely explanation for the remaining NMP was the trapping of solvent molecules within the benzophenone crystal lattice, resulting in a dramatic drop in its vapor pressure and thus inefficient removal. It is likely that the remaining NMP can be removed at a temperature above the melting point of benzophenone; however, some loss of benzophenone due to evaporation would be unavoidable.

The estimated mass loss of benzophenone (after deducting the residual NMP content) was in the range of 2.2−2.4%, which is noticeably less than that in the DMSO trials (9−12%). A possible explanation of this phenomenon is that, in the DMSO trials, samples remained liquid even at the end of the experiments, while at the end of the NMP trials, they were all solid. The room temperature (25.5 °C) vapor pressure value of benzophenone is 1.12 × 10⁻³ and 6.9 × 10⁻⁴ torr for liquid and solid samples, respectively.¹⁴

Spontaneous freezing of the solution being evaporated can result in conditions that are similar to freeze-drying. To explore this possibility, we tested a series of diluted aqueous polyethylene glycol-1000 solutions. Using a setup with a heat-insulated distillation flask (Figure 7, Example 7) afforded a rapid (15−20 min) freezing of the solution during the initial pumping stage, and the lowest measured temperature of the outside bottom surface of the flask was −17 °C. This technique enabled recovery of a water-free polymer in the form of a free-flowing fluffy solid within 7−8 h. In the control tests performed on the same polymer at ambient temperature (as shown in Figure 4), the water evaporation was incomplete even after an extended time of 14 h (5−7 wt % of the water remained), and the polymer was recovered in the form of a waxy mass.

CONCLUSIONS

Solution-phase synthesis workup procedures dealing with isolation of heat-sensitive products may benefit from the ambient-or low-temperature solvent removal method described in this paper. The distillation in this process is driven by the pressure gradient between a distillation flask and a receiver in an evacuated closed system containing a minimal amount of residual noncondensable gas. In principle, the described method is analogous to molecular distillation but with a greatly increased mean free path of a molecule of the distillate vapor. The method’s versatility is exemplified by distillation of dimethylformamide, dimethyl sulfoxide, and N-methylpyrrolidone; 50 mL of each was evaporated at 18−28 °C over the course of 90−140 min when liquid nitrogen was used as a cryogen. The practicality of the method was further demonstrated by the distillation of water at 21 °C into an ice-chilled receiver, which can be useful for both smaller-and larger-scale processes. The utility of the described method is also exemplified by DMSO and NMP solvent removal from solutions of benzophenone, monitored by gravimetric and ¹H NMR methods. Finally, a slight modification of the demonstrated method to yield freeze-drying conditions (by reducing heat flow to the distillation flask) was shown to be applicable for recovery of water-soluble polymers. The procedure is performed in an all-glass apparatus and can therefore be used for the removal of volatile corrosive liquids, including halogenating and acylating agents, solutions of hydrogen halides, and other volatile acids, with minimal risk of equipment damage. This method is recommended for general organic, biological, organometallic, and coordination chemistry solution-phase synthesis routines on a basic or applied research and development scale. Due to the key role the cryogen plays in driving the evaporation of the solvent in the procedure, we propose to name this solvent removal method “cryovap.”

EXPERIMENTAL SECTION

Solvents, benzophenone, and polyethylene glycol-1000 were purchased from ACROS Organics, Alfa Aesar, and Sigma-Aldrich, respectively, and used without further purification. Schlenk and recovery flasks and the distillation bridge were purchased from Chemglass Inc. The condenser/receiver from Figure 2 was ordered from the Chemglass Custom Glass Shop. A Leybold Trivac E2 rotary vane pump was used for evacuation of the demonstrated apparatus. NMR spectra were taken on a Bruker Fourier 300 MHz instrument in CDCl₃. The elemental analyses were performed by Galbraith, Inc.

Example 1.

A 200 mL pear-shaped Schlenk flask was charged with a 10 mm stir bar, 50 mL of DMSO, and 0.5 mL of isooctane. The receiver of the apparatus was charged with 0.5 mL of isooctane, carefully greased, and attached to the Schlenk flask containing DMSO. As the Schlenk flask was held over a working magnetic stirrer, the apparatus was evacuated for 2−3 min. During pumping, the isooctane evaporated and its vapor expelled the residual air within the apparatus. The stopcock was closed, and the system was disconnected from the vacuum line and arranged as shown in Figure 4. The water bath was filled with water and prewarmed to 28 °C (to supply thermal energy for evaporation and to prevent the DMSO from freezing), and the magnetic stirrer was set at high rpm. After about 5 min of stirring, the temperature between the water bath and the solvent was assumed to reach equilibrium. A Dewar flask charged with liquid nitrogen (about one-third of its volume) was adjusted to a position so that only the bottom of the receiver was submerged in the cryogen. Over the course of evaporation, the water bath temperature was maintained between 27 and 28 °C, and the receiver was gradually immersed deeper into the Dewar flask. The solvent in the distillation flask remained liquid at all times, and the distillation was complete in 105 min.

Example 2.

A 200 mL pear-shaped Schlenk flask was charged with a 10 mm stir bar and 50 mL of DMF. The receiver of the apparatus was charged with 0.5 mL of DMF, carefully greased, and attached to the Schlenk flask containing DMF. As the Schlenk flask was held over a working magnetic stirrer, the apparatus was evacuated for 4−5 min. During pumping, a portion of DMF evaporated and its vapor expelled the residual air within the apparatus. The stopcock was closed, and the system was disconnected from the vacuum line and arranged as shown in Figure 4. The water bath was filled with ambient-temperature water (to supply thermal energy for evaporation), and the magnetic stirrer was set at high rpm. A Dewar flask charged with liquid nitrogen (about one-third of its volume) was adjusted to a position so that only the bottom of the receiver was submerged in the cryogen. Over the course of evaporation, the water bath temperature spontaneously dropped to 17−18 °C, so minimal heat was applied to maintain its temperature at ∼25 °C, while the receiver was gradually immersed deeper into the Dewar flask. The distillation was complete in 90 min.

Example 3.

A 200 mL pear-shaped Schlenk flask was charged with a 10 mm stir bar, 50 mL of NMP, and 0.5 mL of isooctane. The receiver of the apparatus was charged with 0.5 mL of isooctane, carefully greased, and attached to the Schlenk flask containing NMP. As the Schlenk flask was held over a working magnetic stirrer, the apparatus was evacuated for 3−4 min. During pumping, the isooctane evaporated and its vapor expelled the residual air within the apparatus. The stopcock was closed, and the system was disconnected from the vacuum line and arranged as shown in Figure 4. The water bath was filled with ambient-temperature water (to supply thermal energy for evaporation), and the magnetic stirrer was set at high rpm. A Dewar flask charged with liquid nitrogen (about one-third of its volume) was adjusted to a position so that only the bottom of the receiver was submerged in the cryogen. Over the course of evaporation, the water bath temperature spontaneously dropped to 19 °C as the receiver was gradually immersed deeper into the Dewar flask. The distillation was complete in 140 min.

Example 4.

A 200 mL pear-shaped Schlenk flask was charged with a 10 mm stir bar and 50 mL of water. The round-bottom flask receiver was charged with 0.5 mL of water, carefully greased, and attached to the Schlenk flask through a distillation bridge. As the Schlenk flask with water was held over a working magnetic stirrer, the apparatus was evacuated for 5 min. During pumping, a portion of water evaporated and its vapor expelled the residual air within the apparatus. The stopcock was closed, and the system was disconnected from the vacuum line and arranged as shown in Figure 5. The water bath temperature was set to 21−22 °C, the magnetic stirrer was set at high rpm, and the ice bucket was charged with ice. The distillation was complete in 3 h.

Example 5.

A 100 mL pear-shaped Schlenk flask was charged with a 10 mm stir bar, 1.000 g of benzophenone, 20 mL of DMSO, and 0.5 mL of isooctane. The receiver of the apparatus was charged with 0.5 mL of isooctane, carefully greased, and attached to the Schlenk flask containing the benzophenone solution. The system was evacuated for ∼2 min, disconnected from the vacuum line, and arranged as shown in Figure 4. The water bath was filled with ambient-temperature water, and the magnetic stirrer was set at high rpm. A Dewar flask charged with liquid nitrogen (about one-third of its volume) was adjusted to a position so that only the bottom of the receiver was submerged in the cryogen. Over the course of evaporation, the water bath temperature spontaneously dropped to 22 °C, and it was readjusted to 25−26 °C. Most of the solvent evaporated in ∼50 min, but the experiment was continued for additional 35 min. Gravimetric control showed that 12% of the benzophenone sample evaporated. The recovered benzophenone contained no residual solvent by ¹H NMR and had a melting point range of 47−47.5 °C.

Example 6.

A 100 mL pear-shaped Schlenk flask was charged with a 10 mm stir bar, 1.000 g of benzophenone, 20 mL of NMP, and 0.5 mL of isooctane. The receiver of the apparatus was charged with 0.5 mL of isooctane, carefully greased, and attached to the Schlenk flask containing the benzophenone solution. The system was evacuated for ∼2 min, disconnected from the vacuum line, and arranged as shown in Figure 4. The water bath was filled with ambient-temperature water, and the magnetic stirrer was set at high rpm. A Dewar flask charged with liquid nitrogen (about one-third of its volume) was adjusted to a position so that only the bottom of the receiver was submerged in the cryogen. Over the course of evaporation, the water bath temperature was maintained at 24−26 °C. Most of the solvent evaporated in ∼50 min, but the experiment was continued for additional 35 min. The recovered benzophenone sample contained 4 mol % of the residual solvent by ¹H NMR (2.2 wt %) and had a melting point range of 41−45 °C.

Example 7.

A 100 mL pear-shaped recovery flask was charged with a 10 mm stir bar, 690 mg of polyethylene glycol-1000, and 10 mL of water. The receiver of the apparatus was charged with 0.5 mL of water, and the joint was carefully greased and attached to the distillation flask containing the polyethylene glycol solution. A K-type thermocouple was taped to the flask’s bottom, the flask was wrapped in a heat-insulating blanket, the magnetic stirrer was set at high rpm, and the system was evacuated for 15 min. By the end of the initial pumping stage, the entire solution was frozen. The evacuation stopcock was closed, and the system was disconnected from the vacuum line and assembled as shown in Figure 7. A Dewar flask charged with liquid nitrogen (about three-fourths of its volume) was adjusted to a position so that most of the receiver was submerged in the cryogen. The lowest temperature of the outside bottom surface of the flask was measured to be −17 °C.

After 7 h, the system was taken apart, and gravimetric control showed a sample weight depression of 1.45% (the final mass was 10 mg lower than the initial), which is most likely attributed to water being present in the original sample. The recovered polymer had a fluffy texture, unlike the initial sample, which was a waxy solid. As a control experiment, the same sample was redissolved in 10 mL of water and freeze-dried under identical conditions. The sample mass depression after the second evaporation was found to be 0.29%, which borders on the accuracy limitation of this method. Elemental analysis calcd (%) for C₄₄H₉₀O₂₃: C 53.53, H 9.19; found: C 53.45, H 9.35.