Artifact-free preparation of biological samples for SEM by optimized water freeze-drying Authors: Hideki Ishida, Naoya Fukuda, Maho Shimada, Kousei Yamamoto, Liudmyla Gaponova, Rina Higuchi, Andrii Kolosiuk, Ryo Hoshina, Mikihiko Arikawa, Yasuhiro Fukuda, MD Shafiq Islam, Terue Harumoto, Yumeng Wan, Toshiki Itoh, Yoko Inai, Asuka Takeishi, Hitoshi Aonuma, Takuto Kikuchi, Shijo Nishigori, Tadashi Maruyama, Kenichi Ikeda, Hideyuki Iriko, Toshinobu Suzaki Published: 2025 ABSTRACT AbstractWater freeze-drying has traditionally been considered an inappropriate drying method for biological samples for scanning electron microscopy (SEM) because ice crystals can potentially form and damage cell structures. Chemical fixation is also considered essential for the SEM observation of soft cells, such as protozoa and animal cells. We succeeded in preparing specimens of freshwater protists for SEM using simple water freeze-drying without chemical fixation. Specifically, a suspension of live cells in water was frozen by contact with a − 80 °C copper block using an outer cylinder that ensured precise planar contact. This significantly reduced artifact formation and increased the success rate from ~ 10% to 20–30%. Transmission electron microscopy (TEM) observations confirmed that intracellular ice crystal formation was significantly suppressed in approximately 5% of Paramecium cells, representing a notable improvement over conventional methods. This method also shortens preparation time to 2–3 h and enables high-resolution SEM imaging of various organisms. Overall, based on these findings, we standardized a simple and low-artifact method using water freeze-drying as a novel and improved technique for SEM sample preparation.Keywords: Electron microscopy, Freeze-drying, Protists, SEM, TEM, Unicellular algaeSubject terms: Biological techniques, Biophysics, Biotechnology, Microbiology Introduction Scanning electron microscopy (SEM) is an essential tool for examining the external morphology of organisms with soft cell surfaces, such as unicellular protists1. However, traditional sample preparation methods involving fixation, dehydration, critical-point drying or t-butanol freeze-drying can significantly cause artifact generation and structural changes in delicate protist cells2–4. To address this limitation, we developed an artifact-minimizing method. This improved water freeze-drying approach successfully yields high-resolution, artifact-free cell morphology, previously difficult to achieve with conventional protocols, with an empirical success rate of 20–30%. Researchers have explored alternative methods that aim to preserve the native state of cells more faithfully. One promising approach is the use of cryogenic techniques, such as freezing samples in liquid nitrogen or in a cryochamber, followed by freeze-drying under vacuum1. Unlike chemical fixation, this method avoids the introduction of exogenous substances and the associated structural changes. Rapid cooling also helps minimize the formation of damaging ice crystals. However, the high vacuum and temperature conditions required for freeze-drying can still cause significant shrinkage and distortion of delicate cells5,6. To address these challenges, we developed an improved water freeze-drying method that preserves the native morphology of various protist cells3,4,7. Our previous reports have indicated that water freeze-drying can greatly reduce the formation of artifacts that occur during fixation and dehydration in some freshwater ciliates. However, it is unknown whether this method can be applied to other organisms. Moreover, there is no theoretical explanation of the effectiveness of this method. Therefore, this paper presents the results obtained by applying this method to ciliates not used in previous reports, non-freshwater protozoa, and even biological samples other than protozoa such as Caenorhabditis elegans, and proposes a standard protocol for this method. Furthermore, based on the results of investigating the changes in temperature and pressure when freezing samples, we discuss the theoretical background of why significant ice crystal formation does not occur using this method. Results Examples of high-resolution SEM images of various protists without chemical fixation Freshwater protists were frozen alive, as shown in Fig. 1, and freeze-dried to obtain satisfactory images. Figure 2 shows SEM images of freshwater ciliates (Sterkiella histriomuscorum, Euplotes aediculatus, and Blepharisma japonicum) without chemical fixation. In all ciliates, the active movement of the cilia (metachronal waves) was momentarily preserved, and no significant deformation of the cell shape was observed. Figure 3 shows SEM images of several freshwater protists apart from ciliates, which were also prepared in an unfixed state. Figure 3A and B show the freshwater alga Micrasterias hardyi, which has well-preserved, fine fibrous hairs growing from the surface of its branches (arrows in Fig. 3B), which may help the cells to float freely in water. Figure 3 also shows SEM images of freshwater amoebae (Amoeba proteus and Mayorella viridis), a species of centrohelid heliozoa, and the freshwater dinoflagellate Ceratium hirundinella. Based on the criteria detailed in the Materials and Methods, the success rate for obtaining artifact-free samples for SEM observation across these various unfixed freshwater protists was empirically determined to be between 20 and 30%. Procedure for water freeze-drying of biological samples for scanning electron microscopy. (A) First, a silicone spacer ring (outer diameter:10 mm; inner diameter:7 mm; thickness: 0.5 mm) is placed on the top surface of an aluminum specimen stub (15 mm in diameter, 10 mm in height). (B) At room temperature, the sample suspension is placed inside the silicone spacer so that the water surface rises slightly above the surface of the spacer ring. (C) The specimen stub was placed in the center of the PVC guide tube (inner diameter: 25 mm). (D) A cylindrical copper block (diameter: 25 mm; height: 30 mm), previously cooled to − 80 °C in deep freezer or on dry ice, is gently placed on top of the specimen stub. (E) The setup is allowed to stand for about 2 min for the temperature of the frozen sample to drop to approximately − 80 °C. (F) The sample stub is then removed from the cooling block. (G) The sample stubs are transferred to a t-butanol freeze-drying apparatus (VFD-21 S, Vacuum Device Co. Ltd., Mito, Japan). In the sample chamber, a copper block (the same one used to freeze the samples) cooled to approximately − 80 °C is placed in the center, and up to four sample stabs are placed in close contact with its sides. The bottom of the freeze-drying apparatus is then set at 0 °C. The ice on the sample stubs will sublimate in about 10 min. However, they are left to dry completely for another hour. After that the temperature of the bottom stage of the freeze-drying apparatus is increased to 30 °C for 1 h. Finally, the chamber is returned to atmospheric pressure, and the sample stubs are removed from the apparatus. Scanning electron microscopy (SEM) images of unfixed freshwater ciliate samples prepared using the water freeze-drying method. (A–C) SEM images of Sterkiella histriomuscorum. A–C indicate the ventral surface of the vegetative cell, a resting cyst, and the adoral band of the membranelles at the anterior end of the vegetative cell, respectively. (D) Euplotes aediculatus cells viewed from the dorsal side. (E) Blepharisma japonicum cells showing adoral membranelles beating in a metachronal manner. Scale bars: A–C: 10 μm; D, E: 20 μm. Scanning electron microscopy images of various freshwater unicellular protists prepared using the water freeze-drying method. (A, B) A and B show a low-magnification image of the unicellular green alga Micrasterias hardyi and a magnified image of the tip of a branch, respectively. The fine fibrous hairs growing from the surface of the branches (arrows) are well-preserved. (C) Image of a large amoeba, Amoeba proteus. (D, E) D and E show a low-magnification image of a whole cell and a magnified image of a scale on the cell surface of a centrohelid heliozoan, respectively. (F) Image of Mayorella viridis, a species of amoebozoan. (G, H) G and H show low- and high-magnification images of the freshwater dinoflagellate Ceratium hirundinella, respectively. The arrows in H indicate openings through which trichocysts are released. Scale bars: A: 20 μm; B: 2 μm; C: 100 μm; D: 2 μm; E: 200 nm; F and G: 10 μm; H: 1 μm. Water freeze-drying of chemically fixed organisms Organisms and tissues that live in environments other than fresh water cannot be freeze-dried in their original state. This is because the components of the external solution may precipitate out. In such cases, it is necessary to chemically fix the sample in advance. After chemical fixation, dehydration with an organic solvent is often performed, but shrinkage of the sample during this process can be problematic. To overcome this, the preferred method is to first freeze the sample to achieve cryo-immobilization. This is followed by stabilization using a chemical fixative solution and washing with pure water. This approach avoids the cell shrinkage typically associated with organic solvent dehydration, allowing subsequent water freeze-drying. Figure 4 shows examples of these: bovine rumen ciliates (A), marine protist, Labyrinthula sp. (B), human red blood cells (C), insect tissues (D and E), and cultured mouse cells (F and G). For all samples, the water freeze-drying method yielded satisfactory results. Scanning electron microscopy images of various biological samples from non-freshwater environments that were frozen, fixed, rinsed with pure water, and then subjected to the water freeze-drying method to minimize artifacts. (A) A bovine rumen ciliate fixed with 2% paraformaldehyde, probably a species of the genus Entodinium. (B) A species of marine Labyrinthula sp. Cells in the exoplasmic net (r) can be identified. (C) Human erythrocytes infected with Plasmodium falciparum NF54 strain (e), and normal erythrocytes (d). (D) A glycerol-treated head section of the trap-jaw ant Odontomachus kuroiwae. (E) A magnified view of the rectangular area in D, showing muscle fibers with characteristic striations. (F) Mouse macrophages spread on a coverslip. (G) A cultured mouse neuroblastoma cell showing many cytoplasmic processes. Scale bars: A and B: 5 μm; C: 2 μm; D: 200 μm; E: 20 μm; F and G: 10 μm. Application of the “double-freezing method” to large organisms For unicellular protists and small animal cells, the method described above gave good results. However, when applying the same method to large multicellular organisms, we observed a marked shrinkage of the organisms. For example, nematodes (C. elegans) formed prominent wrinkles on their body surfaces when freeze-dried in water without the use of a fixative, as shown in Fig. 5A. This shrinkage of the body surface occurred during drying and not freezing. To solve this problem, we developed a double-freeze method targeting large multicellular organisms and verified its effectiveness. As shown in Fig. 5C, it was possible to prepare samples without noticeable shrinkage. Specifically, the samples were frozen once, thawed in a 3% glutaraldehyde solution, rinsed with pure water, and frozen again. This method may increase the potential for artifacts during freezing; however, tissue shrinkage due to dehydration was minimized. Application of the double-freezing method to large organisms. (A, B) When the nematode Caenorhabditis elegans was subjected to freeze-drying in water without fixation in the same manner as in protists, its body surface shrank and developed many wrinkles. (C, D) When unfixed organisms were once water-frozen, thawed in fixing solution (3% glutaraldehyde), washed with water, and freeze-dried again, the body surface was observed to appear smooth. A and C: low-magnification images; B and D: magnified images of the front end (mouth) corresponding to A and C. Scale bars: A: 50 μm; B: 1 μm; C: 50 μm; D: 10 μm. Effect of pressure on structure preservation during water freezing In this method, samples were frozen by applying a cold copper block. Because the pressure applied to the sample could potentially affect the preservation of its morphology during freezing, the pressure applied to the sample was measured using a pressure-sensitive film placed beneath the sample. The morphology of the same sample after freeze-drying was examined with a field emission (FE)-SEM. Three different conditions were used to freeze the samples. The first was to gently place the cooling block on top of the sample (Fig. 6A), the second was to suspend the cooling block 3.5 cm above the sample (Fig. 6B), and the third was to suspend the cooling block 8 cm above the sample (Fig. 6C). For all freezing methods, the morphology and actively beating cilia of Paramecium were found to be well preserved. When the cooling block was gently placed, the pressure was below the limit of measurement (0.6 MPa), except for a small portion, as indicated by the arrow in Fig. 6A. The pressure increased with the height at which the cooling block was dropped; however, when the block was suspended 8 cm above the sample (Fig. 6C), the film coloration was sparse, indicating that the average pressure did not reach 2.5 MPa. These results indicate that the sample pressure does not increase significantly during freezing and that it is not necessary to tap the cooling block against the sample surface at a high speed to freeze the sample. Measurement of the pressure applied to the samples. The pressure applied when the sample freezes was measured using the pressure-measuring film Prescale LLW, which can detect applied pressure in the range of 0.6 to 2.5 MPa. Freezing was performed in three different ways: by gently placing the cooling block on the sample (A) or by dropping the block 3.5 cm (B) or 8 cm (C) above the sample. A suspension of Paramecium multimicronucleatum was placed on top of the pressure measurement film and observed using field emission-scanning electron microscopy after freeze-drying (top row). Pressure can be estimated from the degree of coloration of the pressure measurement film; the results are shown in the bottom row. The schematic on the right-hand side of the figure illustrates the position of the pressure measurement film in the experimental setup. The arrow in Fig. 6A indicates a small area where high pressure was detected, but the average pressure was below the lowest measurable limit of the film (0.6 MPa) in most areas. Cooling curve of pure water To understand how the specimens froze under the conditions used in this study, the cooling profile of the pure water was examined. Temperature measurements were obtained at 25-ms intervals using a small thermocouple positioned at the center of the water sample through a 0.5-mm-thick silicone spacer (Fig. 7). In this experiment, a cooling block was gently placed on top of the sample. At the initial moment (time = 0), the sample stage, which had been set to room temperature (25.0 °C), made contact with a copper block that had been cooled to the temperature of dry ice (–78.5 °C). After an interval of approximately 300 ms, the temperature surpassed 0 °C and decreased at a rate that remained nearly constant at 70 °C/s until an evident inflection point in the graph was attained at 675 ms. At this point, the cooling rate increased abruptly to approximately 380 °C/s. Freezing profile of pure water under the conditions used in this study. Temperature measurements were obtained at 25-ms intervals with a small thermocouple positioned at the center of the sample. The initial phase (Phase 1) and the subsequent phase (Phase 2) were each fitted with a curve using Newton’s law of cooling (Eq. 1, see text). The theoretical curves that provided the best fittings are shown as red dashed lines (Phase 1) and black dashed lines (Phase 2). The graphs replotted using Eq. 2 for the first 0.3 s of each phase are shown in the insets. The proportionality constant r for each phase was estimated from the slope of these replotted graphs and approximated using a straight line. The schematic on the right-hand side of the graph illustrates the position of a small thermocouple placed in the sample water. Upon freezing, the cooling block was gently placed over the sample. The cooling curve shown is a representative profile of multiple repeatable measurements, highlighting the reproducibility of the inflection point and the absence of the thermal jump, which are the key physical phenomena. When water is heated and cooled, the temperature change is usually exponential8. Newton’s law of cooling enables us to describe how temperature changes, despite some limitations9, as illustrated in Eq. (1) below: where T is the temperature of the object (water sample), Tinit is the initial temperature of the object, Tenv is the temperature of the environment (copper cylinder), t is time, and r is a temperature decrement constant that describes the rate at which the object cools (depending on the heat capacity of the object, heat transfer coefficient between the object and environment, and thermal diffusion coefficient of the object). The cooling process is characterized by two distinct phases: phases 1 and 2. The cooling curves for each phase were fitted using Eq. (1). To estimate the temperature decrement constant r, the following equation, which is a variant of the given Eq. (1), can be used: As shown in Fig. 7, the cooling curve data are replotted as graphs of F(t) vs. t. In both phases, the respective graphs were approximated as straight lines, confirming an exponential decrease in temperature. Here, we used − 75 °C as Tenv, not the − 78.5 °C temperature of dry ice, to achieve better linearity. This is likely because the sample was in thermal contact with the 25 °C sample stage as well as the cooled copper block. In each phase, the temperature decrement constant r was calculated as the slope of the linear regression curve based on the first 0.3 s of the data; its value was 0.79 in phase 1 and 4.17 in phase 2, more than five times greater than in phase 1. This difference can be explained by the fact that the specific heat capacity of ice is about half that of water, and its thermal conductivity is much higher. TEM after water freezing Live P. multimicronucleatum cells frozen using the method described in this paper were examined using TEM after freeze substitution. The cytoplasm of most cells exhibited spongy artifacts; however, a small percentage (approximately 5%) of cells were observed to be in excellent condition, even after freezing. As shown in Fig. 8, such cells showed no significant ice crystal formation at least approximately 10 μm inward from the cell surface. Moreover, the tubular cristae inside the mitochondria and the finely furrowed surface structures of the trichocysts were also well-preserved (Fig. 8B and C). Ultrathin section images of freeze-substituted P. multimicronucleatum after water freezing. (A) Low-magnification image of the cell surface layer. There is no noticeable sign of ice crystal formation from the cell surface down to about 10 μm. (B, C) Magnified images of the areas indicated by “b” and “c” in Fig. 8A. Tubular cristae, a structure characteristic of ciliates, can be observed inside the mitochondria (m) in (B) The furrowed structure on the surface of trichocyst (t) can be seen in (C) The alveoli (a) just below the cell surface are also well-preserved. Scale bars: A: 2 μm; B: 0.5 μm; C: 1 μm. Discussion We have previously reported satisfactory results for the water freeze-drying method using various ciliates3,4,7. The present study also showed good results in other ciliates, microalgae, and protozoa, as well as in small animals and cultured cells. In particular, we have shown that samples of unfixed freshwater organisms can be prepared and that the water freeze-drying method can be used even when the external environment is not freshwater, especially if the samples were chemically fixed in advance. The conditions for reducing ice crystal formation upon water freezing include either rapid freezing or extremely high pressures10. Because the environmental conditions used in this study were open-type, it is unlikely that high pressure (> 200 MPa) was reached to suppress ice crystal formation. Measurements using a pressure-sensitive film supported this hypothesis (Fig. 6). Slamming the cooling block into the sample suspension at high speed seems to produce artifacts, such as unnecessary physical shock to the cells before freezing, which causes the partial desorption of cilia in Paramecium. These results indicate that gently bringing the cooling block into contact with the sample was the best freezing method. The minimal variation in macroscopically detected pressure, even when comparing a ‘gently placed’ block to one ‘dropped from 8 cm’ (Fig. 6), confirms that the procedure is robust against slight operator variability in applying the cooling block. When water is supercooled at atmospheric pressure, its solidification is initiated via partial freezing, resulting in a rapid rise in temperature to 0 °C. This phenomenon is attributed to the release of latent heat. A temperature of 0 °C represents the thermal equilibrium temperature at which ice and water coexist at atmospheric pressure. Subsequent solidification occurs while maintaining this temperature, and once all water turns into ice, the temperature begins to decrease. In contrast, the temperature change observed in this study indicates that the solidification process occurred without any temperature jump or subsequent temperature plateau. Measurements of the cooling profile of water showed that, in the center of the sample chamber, water showed a temperature drop of − 70 °C/s in the first phase, passing 0 °C in approximately 300 ms. Although this cooling rate cannot be called “rapid cooling,” as it is commonly referred to, the cooling rate at the moment when the sample cells are frozen may actually be much faster due to the existence of supercooling phenomena, as suggested by the presence of an inflection point in the cooling curve at around − 20 °C (Fig. 7). If the sample water freezes without exhibiting supercooling, the inflection point would exist at the point where the temperature of the sample water reaches 0 °C. However, actual measurements show that temperatures are decreasing at the same rate as they pass through 0 °C. This indicates that the sample water may have supercooled and become rapidly frozen when it reached − 20 °C. The most notable feature of our results is the absence of a temperature jump and subsequent temperature plateaus, which are generally observed when supercooled water freezes. The differences in behaviour may be due to the confinement of water in a small space during the experiment. As shown in Fig. 7, in such a confined space, rapid solidification may have resulted in an instantaneous high pressure owing to volume expansion. Under high pressure, expansion generates significant mechanical force, and a portion of the latent heat is utilized in this process. Furthermore, the freezing point on the water-ice phase boundary under high pressure is much lower than 0°C11, the freezing point at atmospheric pressure. This freezing point lowering is also considered one of the factors preventing temperature jumps. To decrease the freezing point of water to − 20 °C, approximately 200 MPa of pressure is needed11. The ice formed under these pressure and temperature conditions was considered to be in the Ice IV phase, where ice crystal formation was highly suppressed. We acknowledge that due to technical limitations (specifically the size and placement constraints of the current sensors), simultaneous measurement of temperature and pressure at the exact micro-location was not feasible in this experimental setup. This mechanism, which is similar to that of a high-pressure freezer12, may have led to the good drying results observed in this study. The internal structure of the cells was examined using TEM, and a small percentage of cells with good freezing conditions were observed. In such cells, the preservation of the surface layer structure was good and almost equivalent to that prepared using the metal contact method of rapid cooling to the liquid nitrogen temperature, as shown by Song and Suzaki13. This indicates that the formation of ice crystals during freezing was effectively suppressed, at least locally. The percentage of cells with well-preserved internal structure was approximately 5%, which was very low compared to the percentage of cells with good observations using SEM, which was about 20–30%. This may be attributed to the fact that ice crystal formation is more likely to be promoted when the intracellular cytoplasm is frozen compared to when pure water freezes outside the cell. Because of its low impurity content, pure water can easily maintain a supercooled state and remain liquid at low temperatures14. However, since good results were obtained in the present study, this freezing method may be useful for observing intracellular structures using state-of-the-art rapid-freezing equipment. The water freeze-drying method introduced here was first used by Ishida et al.7for the ciliate Spirostomum ambiguum and was introduced as the only effective method to instantaneously stop cell movement. Subsequently, Ishida et al.3,4 reported satisfactory results for several other ciliates. The current study further showed that this method can be applied to cells and tissues of protists and animals other than ciliates. The greatest advantage of the water freeze-drying method is that it minimizes the production of artifacts caused by chemical fixation and dehydration. Unfixed SEM observations are possible for organisms living in freshwater environments. Dehydration with organic solvents, such as ethanol or acetone, is considered essential for sample preparation in conventional SEM and TEM; however, dehydration has been shown to cause significant cell shrinkage in both SEM3and TEM13. The water freeze-drying method virtually eliminates such artifacts from the sample, enabling the observation of an accurate three-dimensional shape. In the double-freezing method, applying the fixative (glutaraldehyde) to the cryo-immobilized structure after rapid freezing, rather than to a living organism, avoids the structural reorganization typically caused by conventional chemical fixation. This feature shares a mechanistic advantage with high-end cryogenic techniques such as High-Pressure Freezing followed by Freeze-Substitution, thus further minimizing potential fixation artifacts. In addition, this method does not require the use of toxic chemicals as fixatives (especially for protozoa living in freshwater environments), and the only equipment required is for vacuum evacuation. Another significant advantage of this method is that the time from sample freezing to the completion of drying was short (only 2–3 h). However, this drying method has several limitations. The most important limitation of this method was that not all samples showed good drying results. We previously reported that the success rate of this method was approximately 10%4; however, with the modification presented in this paper, the success rate increased to 20–30%. Overcoming this problem is a major goal of future investigations. Conclusion In this study, we demonstrated that water freeze-drying is an excellent sample preparation method for the observation of various biological samples using SEM. The application of the proposed standard experimental procedure revealed no substantial impairment in either the external morphology or the internal microstructure of the cells that can be attributable to ice crystal formation. Furthermore, an examination of the pressure and temperature changes imposed on the sample during the freezing process suggested that rapid freezing occurred due to supercooling, resulting in the suppression of ice crystal formation. Materials and methods Tools and materials needed for water freeze-drying The following items are required for water freeze-drying: Specimen stub for SEM (15 mm diameter, 10 mm height). Specimen stub for SEM (15 mm diameter, 10 mm height). Silicone rubber sheet (≤ 0.5 mm thick). Silicone rubber sheet (≤ 0.5 mm thick). Drilling punch (10 mm and 7 mm). Drilling punch (10 mm and 7 mm). Conductive, double-sided carbon sticky tape. Conductive, double-sided carbon sticky tape. Copper cooling block (25 mm diameter, 30 mm height) with a mirror finish on top. Copper cooling block (25 mm diameter, 30 mm height) with a mirror finish on top. Polyvinyl chloride (PVC) guide tube (25 + α mm inner diameter, 30–50 mm height). Polyvinyl chloride (PVC) guide tube (25 + α mm inner diameter, 30–50 mm height). Freeze-drying apparatus. Freeze-drying apparatus. Dry ice (–78.9 °C) or deep freezer (–80 °C). Dry ice (–78.9 °C) or deep freezer (–80 °C). Water freeze-drying procedure A ring shape was punched out from the silicone sheet using drilling punches (Fig. 1A). A ring shape was punched out from the silicone sheet using drilling punches (Fig. 1A). A piece of double-sided carbon adhesive tape was placed on the specimen stub, and a silicone spacer ring was placed on top of it (Fig. 1A). A piece of double-sided carbon adhesive tape was placed on the specimen stub, and a silicone spacer ring was placed on top of it (Fig. 1A). The inside of the spacer was filled with the sample suspension, and the water surface was adjusted until it rose just a little (Fig. 1B). The inside of the spacer was filled with the sample suspension, and the water surface was adjusted until it rose just a little (Fig. 1B). The specimen stub was placed in the center of the PVC guide tube (Fig. 1C). The specimen stub was placed in the center of the PVC guide tube (Fig. 1C). The cooling block cooled to − 80 °C was gently placed along the guide tube on the surface of the specimen stub or was allowed to hang freely from about 30 mm above it (Fig. 1D). The cooling block cooled to − 80 °C was gently placed along the guide tube on the surface of the specimen stub or was allowed to hang freely from about 30 mm above it (Fig. 1D). The sample was left for about 2 min to freeze the specimen completely (Fig. 1E). The sample was left for about 2 min to freeze the specimen completely (Fig. 1E). The sample stub was then removed from the cooling block. If the sample adhered to the cooling block, sample preparation was considered a failure; steps 1 to 6 above were then repeated (Fig. 1F). The sample stub was then removed from the cooling block. If the sample adhered to the cooling block, sample preparation was considered a failure; steps 1 to 6 above were then repeated (Fig. 1F). The frozen specimen stubs were then placed in the freeze-dryer while maintaining the temperature at − 80 °C (Fig. 1G). The frozen specimen stubs were then placed in the freeze-dryer while maintaining the temperature at − 80 °C (Fig. 1G). The stage temperature of the freeze-dryer was pre-set to 0 °C or lower. The stage temperature of the freeze-dryer was pre-set to 0 °C or lower. An − 80 °C cooling block was placed at the center of the stage. The specimen stub was then placed in close contact with its sides. An − 80 °C cooling block was placed at the center of the stage. The specimen stub was then placed in close contact with its sides. Next, the freeze-dryer was evacuated and allowed to stand for 1 h. The ice in the sample should have sublimated completely in approximately 10 min. Next, the freeze-dryer was evacuated and allowed to stand for 1 h. The ice in the sample should have sublimated completely in approximately 10 min. The stage temperature was then raised to 30 °C and let stand for another hour. The stage temperature was then raised to 30 °C and let stand for another hour. Finally, the sample was removed from the freeze-dryer, a metal coating (OsO4) was applied, and the samples were observed with SEM. Finally, the sample was removed from the freeze-dryer, a metal coating (OsO4) was applied, and the samples were observed with SEM. Important tips for successful water freeze-drying The thickness of the silicone spacer should be 0.5 mm (thinner may be better; more than 1 mm will result in uneven freezing). The thickness of the silicone spacer should be 0.5 mm (thinner may be better; more than 1 mm will result in uneven freezing). The cooling block temperature of − 80 °C is optimal. Higher or lower temperatures do not produce good results. The cooling block temperature of − 80 °C is optimal. Higher or lower temperatures do not produce good results. The cooling block should be placed gently on top of the specimen stub. The cooling block should be placed gently on top of the specimen stub. Water sublimation should be done slowly (the initial temperature should be − 80 °C, and sublimation may take about 10 min). Water sublimation should be done slowly (the initial temperature should be − 80 °C, and sublimation may take about 10 min). Evaluation of success rate and statistical details The success rates of the specimens presented in this paper were calculated empirically based on the total number of cells observed (N value) and the number of independent replicates used. The criteria for judging a specimen as “successful” were defined as being artifact-free, meaning no cell deformation or damage (e.g., deciliation) was observed. The overall success rate for scanning electron microscopy (SEM) observation was 20–30%, and the success rate for transmission electron microscopy (TEM) observation was 5%. For example, in experiments using Paramecium (P. multimicronucleatum), the successful rate for obtaining favorable specimens was 16.8% (n = 286), and the criteria were defined as no cell deformation and no deciliation (loss of cilia). Protists The Sterkiella histriomuscorum F01 strain was originally collected from a pond in Fukui City, Japan (36°04’26.4"N 136°10’22.2"E). Extracted DNA was subjected to PCR using the universal primers SR1 (5’ 18 S rDNA)/SR12 (3’). The PCR products were directly sequenced (LC890396), and the BLAST results showed 100% identity to the S. histriomuscorum 17 S ribosomal RNA gene (FJ545743.1). The EaKb1 strain of Euplotes aediculatus was collected in a pond inside Kobe University, Japan (34°43’32.8"N 135°14’07.6"E). Blepharisma japonicum (strain R1072, mating type I) was cultured as described by Kobayashi et al.15. The unicellular green alga Micrasterias hardyi was collected in Lake Biwa, Japan (35°06’33.2"N 135°55’14.7"E). Amoeba sp. was collected from the Kamo River (35°01’45.9"N 135°46’19.5"E) and cultured on brown rice grains as a nutrient source. A centrohelid heliozoan F02 was collected from the same pond as S. histriomuscorum in Fukui, Japan. The 18 S rDNA sequence (LC890397) indicated that this species is related to the genera Raphidiophrys and Choanocystis. Mayorella viridis strain NS-1 was collected from a freshwater lake in Tochigi, Japan (37°06’42.2” N 139°56’07.9” E) and cultured in mineral water. The freshwater dinoflagellate Ceratium hirundinella (NIES-4632) was originally collected in Lake Ashino-ko, Kanagawa, Japan (35°11’51.4"N 139°01’30.3"E). It was grown in a culture medium containing 0.01% KW21 (Daiichi Seimo Co. Ltd., Kumamoto, Japan) in mineral water (Crystal Geyser, Ohtsuka Pharmaceutical Co. Ltd., Japan) containing 1% boiled natural seawater at room temperature, 18–23 °C, under 4.2 W of white LED light illumination (20–80 µmol·m− 2·s− 1) with a 14-h:10-h light-dark cycle. Labyrinthula sp. was collected from the coast of Sumoto City, Hyogo Prefecture, Japan (34°16’26.3"N 134°57’09.6"E), and cultured in sterile seawater at 20 °C with diatoms as a food source. Animal specimens Wild-type N2 Caenorhabditis elegans were cultured on NGM agar plates (0.25% peptone, 50 mM NaCl, 25 mM KPO4, 5 µg/mL cholesterol, 1 mM CaCl2, 1 mM MgCl2, 1.7% agar) with Escherichia coli OP50 as a food source. C. elegans was washed twice with M9 buffer (20 mM KH2PO4, 40 mM Na2HPO4, 8.5 mM NaCl, 20 mM NH4Cl) and resuspended in distilled water. Adult C. elegans were used for SEM imaging. C. elegans N2 strain and E. coli OP50 were obtained from the Caenorhabditis Genetics Center (CGC). Workers of the trap-jaw ant Odontomachus kuroiwae (Matsumura 1912) were used in this study. Briefly, ant colonies were collected from Okinawa and reared as described by Aonuma et al.16. After anesthesia with CO₂ gas, the ant’s head was cut off, and the distal part was mounted on the tip of a plastic tube with a diameter of 2 mm using UV-curable resin (Bulufixx PW, BLUFIXX, Ellwangen, Germany). The antennae, mandibles, and mouthparts were removed from the head using fine forceps. The samples were then mounted on a micro slicer (THK, Kenis, Osaka, Japan) and sliced into 200-µm-thick sections. Cross-sections of the head at 600 μm and 800 μm from the rostral end were stored in 60% glycerol at 4 °C and then fixed with 3% glutaraldehyde. Human erythrocytes infected with the Plasmodium falciparum NF54 strain were fixed with a fixative containing 1% paraformaldehyde and 0.2% glutaraldehyde. RAW 264.7 cells, a murine macrophage cell line, were purchased from the American Type Culture Collection. The cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; 08458-16, Nacalai Tesque, Tokyo, Japan) supplemented with 10% fetal bovine serum (F7524, Sigma-Aldrich, St. Louis, MO, USA) under a 5% CO2 atmosphere at 37 °C. Mouse neuroblastoma (Neuro-2a) cells were obtained from Health Science Research Resources Bank and also cultured in DMEM supplemented with 10% fetal calf serum, 100 µg/mL streptomycin, and 100 U/mL penicillin. The mouse cells were fixed with 3% glutaraldehyde for 30 min at room temperature and washed with pure water before use. Rumen fluid collection and prefixation Bovine rumen fluid was collected from Holstein cows, which are bred in the Kawatabi Field Science Center (Tohoku University, Miyagi, Japan), using a stomach tube. An equal volume of a 4% paraformaldehyde solution was then added to the collected ruminal fluid for prefixation. The prefixed ruminal fluid was stored at 4 °C until further processing. Ruminal fluid collection was carried out by the authors at the Kawatabi Field Science Center. This procedure for ruminal fluid collection was approved by the Institutional Animal Care and Use Committee of Tohoku University (Approval number: 2019AgA-033). This study was conducted and described in accordance with the ARRIVE guidelines, and it was also conducted in accordance with the relevant guidelines and regulations. Temperature measurement The temperature of the deionized distilled water (pure water) was measured using homemade chromel–alumel thermocouples comprising wires of equal thickness (25 μm each) with a Keithley 2182 nanovoltmeter (Keithley, USA) at 25-ms intervals. Calibration of the measurement temperatures was performed at 0 °C (ice water), − 78.5 °C (dry ice), and − 195.8 °C (liquid nitrogen). Pressure measurement An LLW two-sheet-type Fuji Prescale film (Fuji Photo Film Co., Ltd., Tokyo, Japan) was placed immediately below the sample suspension to measure the pressure generated during the freezing process, which was estimated using the color samples provided by the manufacturer. Water freeze-drying without chemical fixation The freshwater organisms (protists and unicellular algae) were lightly rinsed with pure water and freeze-dried using the method shown in Fig. 1. The method used was essentially the same as that previously reported4, but this time, a PVC outer casing was used as a guiding structure when freezing. This made it possible to precisely adhere the top surface of the specimen to the bottom surface of the cooling block and drop the cooling block at a constant speed. The use of an outer cylinder also maintains a low sample temperature during cooling, thus dramatically improving the success rate of sample preparation. Water freeze-drying after chemical fixation Mouse Neuro2A neuroblastoma cells were cultured on glass coverslips. To ensure cryo-immobilization before chemical stabilization, the cells were first frozen using the cold copper block method. The frozen sample was then fixed by rapid immersion in a 4% paraformaldehyde solution (or similar chemical fixative) to stabilize the structure. The samples were subsequently washed with pure water and freeze-dried using the method described above. Double-freezing procedure for the nematode C. elegans Owing to their robust outer coats, centipedes and crustaceans may survive for some time after exposure to chemical fixatives. For example, the nematode C. elegans can move around in 3% glutaraldehyde for approximately 30 min. This suggests that rapid freezing is more effective than chemical fixation in rapidly stopping the movement of these organisms and preserving their morphology. Since the freeze-drying of unfixed C. elegans resulted in significant shrinkage of the organism, a “double-freezing procedure” combining freezing and chemical fixation was devised in this study. Transmission electron microscopy (TEM) of freeze-substituted P. multimicronucleatum To investigate the extent of ice crystal formation in samples frozen using the method described in this paper, live P. multimicronucleatum cells were frozen in water and freeze-dried in acetone containing 1% osmium tetroxide. After resin embedding, ultrathin sections were prepared and observed under a Hitachi H-7100 transmission electron microscope (Hitachi, Tokyo, Japan) at 75 kV. Samples were prepared as described by Song et al.17. Acknowledgements We thank the Japanese Red Cross Society for providing human erythrocytes. We would like to thank Editage (www.editage.jp) for English language editing. Author contributions H. Ishida and T. Suzaki conceived the project; H. Ishida, T. Suzaki, N. Fukuda, M. Shimada, K. Yamamoto, S. Nishigori designed all research; H. Ishida, N. Fukuda, K. Yamamoto, T. Suzaki, R. Hoshina performed the experiments and descriptions; L. Gaponova, R. Higuchi, A. Kolosiuk, M. Arikawa, Y. Fukuda, M D S. Islam, T. Harumoto, Y. Wan, T. Itoh, Y. Inai, A. Takeishi, H. Aonuma, T. Kikuchi, T. Maruyama, K. Ikeda, H. Iriko prepared biological specimens; H. Ishida, T. Suzaki, S. Nishigori, A. Kolosiuk analyzed the data; H. Ishida, T. Suzaki wrote the article. Funding This work was supported by JSPS KAKENHI Grant Number JP21K19168, JP22K06377, JP24KJ1708. Data availability The datasets generated and/or analysed during the current study are available in the GenBank/EMBL/DDBJ repositories, accession numbers LC890396 and LC890397. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The datasets generated and/or analysed during the current study are available in the GenBank/EMBL/DDBJ repositories, accession numbers LC890396 and LC890397.