Sterilization using high-pressure carbon dioxide

J. of Supercritical Fluids 38 (2006) 354–372 Sterilization using high-pressure carbon dioxide Jian Zhang a , Thomas A. Davis a , Michael A. Matthews a,∗ , Michael J. Drews b , Martine LaBerge c , Yuehuei H. An d a Department of Chemical Engineering, University of South Carolina, Columbia, SC 29208, USA School of Materials Science and Engineering, Clemson University, Clemson, SC 29634, USA c Department of Bioengineering, Clemson University, Clemson, SC 29634, USA d Orthopaedic Research Laboratories, Medical University of South Carolina, Charleston, SC 29403, USA b Received 26 January 2005; received in revised form 7 May 2005; accepted 31 May 2005 Abstract Sterility is required for medical devices use in invasive medical procedures, and for some situations in the food industry. Sterilization of heatsensitive or porous materials or devices, such as endoscopes, porous implants, liquid foodstuff, and liquid medicine, poses a challenge to current technologies. There has been a steady interest in using high-pressure carbon dioxide as a process medium for new sterilization technology. Among the potential advantages are that CO2 may sterilize at low temperatures. This paper is a review of the technical and patent literature, including analysis of the microorganisms studied, important operating parameters, and deactivation mechanisms. The current research status and challenges are summarized at the end of this paper. © 2006 Elsevier B.V. All rights reserved. Keywords: Sterilization; High-pressure carbon dioxide 1. Introduction Sterilization of implants is crucial to prevent infecting patients. In the United States, over 600,000 arthroplasties are performed each year [1], of which 0.6–2.3% result in infection. This can cause physical injury or even death to the patients [2]. Other widely used medical devices, such as endoscopes, can also cause infection if not properly sterilized between uses. Disinfection of heat-sensitive biomaterials, especially polymers, presents a challenge to current sterilization technology. In medical practice the standard sterilization methods include steam, gammairradiation, ethylene oxide, and hydrogen peroxide sterilization [3,4]. Each method has drawbacks in certain applications, as summarized in Table 1. Steam sterilization is the most common technique because of its low cost and effectiveness. However, steam sterilization operates at 121 ◦ C so heat-sensitive materials will be damaged or destroyed [4]. Additionally, steam sterilization may deposit an oxide layer onto metallic devices, which decreases the biocompatibility of the treated implants [5]. Unlike steam sterilization, ␥-irradiation and ethylene oxide sterilization ∗ Corresponding author. Tel.: +1 803 777 0556; fax: +1 803 777 8265. E-mail address: matthews@engr.sc.edu (M.A. Matthews). 0896-8446/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.supflu.2005.05.005 can be applied to heat-labile materials. However, ␥-irradiation may cause changes in shear and tensile strength, elastic modulus, and transparency of medical polymers [4]. For example, longlived free radicals generated by ␥-irradiation cause progressive oxidation, breaking of polymer chains, and deterioration of mechanical properties in ultra high molecular weight polyethylene (UHMWPE) [6]. Ethylene oxide is flammable and toxic, is a known carcinogen and can cause hemolysis [4]. Ethylene oxide sterilization can also chemically alter material properties, such as the molecular weight of biodegradable polymers [4]. Because of limitations with current sterilization techniques, the next generations of polymeric medical devices and heatsensitive biomaterials demand new sterilization methods [7]. The biocidal effects of high-pressure carbon dioxide (CO2 ) have been quantified on various species of bacteria and have been summarized elsewhere by Spilimbergo and Bertucco (2003) [8]. Using CO2 as a sterilant has several potential benefits. First, CO2 is not flammable and is non-toxic; the chief hazard in its use is asphyxiation. Unlike ethylene oxide, CO2 requires no special handling or ventilation, and leaves no toxic residues. Second, CO2 is inert in most situations so it does not react with polymers, which alleviates the aging problem caused by ␥-irradiation. Next, CO2 has a low critical temperature (31.1 ◦ C) [9]. This is only slightly above room temperature, J. Zhang et al. / J. of Supercritical Fluids 38 (2006) 354–372 355 Table 1 Advantages and limitations of sterilization methods Effects Steam EtO ␥-Irradiation UV CO2 Sterilization of inside surfaces Surface cleaning Surface contamination, decreased biocompatibility Effects on mechanical properties of polymers Yes No Yes Yes Yes No Yes Yes Yes No No Yes No No No No Yes Yes No No so thermal degradation is not a problem when a process is operated around the critical temperature. Because of these benefits CO2 has been proposed for use in other biomaterial applications such as incorporating bioactive ingredients into biodegradable polymers [10] and producing enzyme particles [11]. Moreover, in the supercritical state, CO2 has low viscosity (3–7 × 10−5 N s m−2 ) and zero surface tension [9], so it can quickly penetrate complex structures and porous materials. Finally, CO2 is inexpensive and readily available, which makes switching to CO2 -based sterilization economically feasible. A preliminary economic analysis has been conducted to estimate the feasibility of this novel technique [3]. Taking into account the low operating temperature and absence of toxic residue, CO2 -based sterilization may be superior to steam and ethylene oxide sterilization for certain applications, and be competitive with hydrogen peroxide sterilization. In short, CO2 is inexpensive, non-toxic, non-flammable, physiologically safe, with a low critical temperature, therefore, a high-pressure CO2 sterilization technique could be an option for heat-sensitive and/or porous medical devices and biomaterials sterilization. ization. Bacterial endospores (Bacillus subtilis and Geobacillus stearothermophilus spores) were the test organisms. However, even after adding 0.5 wt.% acetic acid or 2 wt.% ethanol as an entrainer, only 57 and 38% of G. stearothermophilus spores were destroyed, respectively [15]. These results demonstrate that bacterial spores are highly resistant to high-pressure CO2 treatment. Another major contribution from these studies was the postulation of a physiological deactivation mechanism, suggesting that cell deactivation was due to enzyme deactivation and extraction of cytoplasmic materials by high-pressure CO2 [15]. In subsequent years more than 50 journal papers and research reports have been published on high-pressure CO2 sterilization. In those publications, 36 species of microorganisms have been studied (Tables 3–6) under a wide range of operating conditions (Table 7). A number of experimental techniques have been adopted to characterize morphological or structural changes induced by high-pressure CO2 treatment (Table 10), and new theories of the deactivation mechanisms have been proposed (Table 9). 3. Gases evaluated for sterilization 2. Early work In 1951, Fraser proposed a novel technique to collect the contents of Escherichia coli by bursting cells in liquid culture with a sudden release of pressurized (1.7–6.2 MPa) gases (Ar, N2 , N2 O, and CO2 ) [12]. E. coli deactivation rates between 95 and 99% were achieved with 3.4 MPa of CO2 . This level of deactivation is acceptable for collecting cell contents. However, in terms of sterilization, 99% deactivation is only approximately a 2-log reduction, far below the >6-log reduction required for medical sterilization [4]. The concept of treating microorganisms with fast depressurization of high-pressure gases was further examined by Foster et al. in 1962 [13]. Six species of microorganisms were processed with nitrogen at 12 MPa in a specially designed apparatus with instantaneous pressure release. The experiments achieved up to 58.9% cell rupture. Electron microscope images of treated cells were used to support Fraser’s theory of cell rupture. In the late 1980’s, strict regulations on ethylene oxide and radiation sterilization in Japan inspired several Japanese researchers to explore the use of high-pressure gas treatment as an alternative for sterilizing biological products (e.g. plasma powder) [14] and heat-labile compounds [15], and for preserving foods [16]. These studies focused on CO2 because of the benefits mentioned above. Both entrainers (or, co-solvents) and water content, were examined by Kamihira et al. [15] and Taniguchi et al. [14], and were shown to be beneficial for steril- Though the majority of publications have focused on highpressure CO2 , some studies exploited other gases, including nitrogen, nitrous oxide, ethylene, argon, and tetrafluoroethane (TFE) (Table 2). Fraser treated E. coli with Ar, N2 , N2 O, and CO2 at 37 ◦ C and a pressure between 1.7 and 6.2 MPa for 5 min and found that the other three gases were less effective than CO2 [12]. Wei et al. showed that CO2 treatment at 35 ◦ C, 5.5–6.2 MPa, for 2 h completely deactivated L. monocytogenes, while N2 treatment at the same condition had no effect [17]. Similar results have been reported by Nakamura et al. [16], Enomoto et al. [18], and Dillow et al. [4]. Dillow et al. concluded that using N2 far from the critical point (Tc = −147 ◦ C, Pc = 3.39 MPa) sacrificed the unique properties (gas-like diffusivity and liquid-like density) of a supercritical fluid. They found CO2 to be much more effective than N2 at the same conditions of temperature and pressure [4]. Not only is the vicinity to a critical point important, but also the chemical properties of the gas are important. Dillow et al. tested sterilization effects of TFE, which has a critical point (Tc = 328 K, Pc = 4.06 MPa) similar to CO2 (Tc = 304.13 K, Pc = 7.38 MPa), but different chemical properties (dipole moment, DCO2 = 0 D, DTFE = 1.80 ± 0.22 D; solubility parameter, δCO2 = 7.0, δTFE = 13.6) [4]. At the same reduced temperature and reduced pressure as successful CO2 experiments, no reduction of viable cells was observed after TFE treatment. Another gas having a biocidal effect is N2 O. Castor and Hong [19] reported that supercritical N2 O is more effective than N2 356 J. Zhang et al. / J. of Supercritical Fluids 38 (2006) 354–372 Table 2 Properties of gases used for sterilization Gas Tc a (K) Pc a (MPa) Dipole momenta (D) δ at 25 ◦ C (MPa1/2 ) Solubility in H2 O at 25 ◦ C and 1 atm (mol/mol) Ref. Carbon Dioxide Argon Nitrogen Nitrous oxide TFE 304.13 150.87 126.21 309.57 328c 7.375 4.898 3.39 7.255 4.065 0 0 0 0.16 1.80 12.3b 10.9b 5.3 b NA 13.6c 6.15 × 10−4a 2.519 × 10−5a 1.183 × 10−5a 4.376 × 10−4a 2.646 × 10−4d Most of the references [12,18] [4,13,16–19,28,43,51,53] [12,18,19] [4] a b c d CRC Handbook of Chemistry and Physics (84th ed.). CRC Handbook of Solubility Parameters and Other Cohesion Parameters (2nd ed.). Dillow et al. [4]. MSDS of 1,1,1,2-tetrafluoroethane. in recovering nucleic acids from vegetative bacteria and yeast. This is possibly because the high density and low polarity of supercritical N2 O could favor solubilizing lipids and hydrophobic compounds in the cell wall and the cytoplasmic membrane. Enomoto et al. reported biocidal effects of N2 O on yeast cells at 4 MPa, 40 ◦ C after 4 h exposure. However, N2 O failed to deactivate B. megaterium spores at 6 MPa, 60 ◦ C even after 24 h [18]. The biocidal effect of N2 O may be a result of proximity to the critical point and its solubility in water. The critical parameters of N2 O are very close to those of CO2 ; N2 O only has a small dipole moment, while CO2 has a zero dipole moment; both have a comparatively high solubility in water. 4. Microorganisms used in high-pressure CO2 treatment ber of publications. This is in accordance with the fact that food poisoning [20] and foodborne diseases [21] are caused by nonsporulating vegetative bacteria. The objectives of those studies were either to preserve foods [22], to improve product quality [22], or to recover bioproducts [23]. Different target microorganisms have been studied, depending on the practical application. In the area of food spoilage, pathogenic bacteria such as Listeria monocytogenes [17,24–27], Staphylococcus aureas [4,15,20,28] and Salmonella typhimurium [17,21,29] are of interest. Hong et al. treated Lactobacillus plantarum in order to solve the problem of over-acidification caused by proliferation of L. plantarum in the late stage of kimchi production [30,31]. Bacterial spores are the least frequently studied type of organism. Fewer than 20% of the studies are dedicated to treatment of spores, possibly because spore deactivation is not required for food preservation. 4.1. Overview 4.2. Treatment of vegetative bacteria The papers reviewed herein cover 12 species of gram-positive bacteria, 10 species of gram-negative bacteria, spores of eight species of bacteria, and eight species of fungi (fungus and fungal spores). Fig. 1 shows the distributions of the number of species and the number of publications in each category. Studies on vegetative bacteria account for approximately 60% of all the studies reported, in terms of both the number of species and the num- Traditionally, bacteria are categorized into two major groups, gram-positive and -negative bacteria, according to their responses to the gram stain [32]. The different response to the gram stain derives from differing peptidoglycan content. Gram-positive cell walls are simple in structure, but have thick peptidoglycan layers (10–20 layers thick, as much as 90% of the cell wall), which make the cell walls strong and robust. How- Fig. 1. Distribution of the number of species (gram-positive bacteria, gram-negative bacteria, bacterial spores, and fungi) and the number of publications by species. J. Zhang et al. / J. of Supercritical Fluids 38 (2006) 354–372 357 Table 3 Gram-positive vegetative bacteria species treated with dense-phase carbon dioxide Species Researcher (year) Geobacillus stearothermophilus Roskey and Sikes (1994) Sikes and Martin (1994) Bacillus subtilis Elvassore et al. (2000) Spilimbergo et al. (2002) Parton et al. (2003) Brochothrix thermosphacta Enterococcus faecalis Erkmen (2000) Debs-Louka et al. (1999) Erkmen (2000) Lactobacillus brevis Ishikawa et al. (1995) Shimoda et al. (1998) Lactobacillus casei Lactobacillus plantarum Haas et al. (1989) Smelt and Rijke (1992) Hong et al. (1999) Hong and Pyun (1999) Hong and Pyun (2001) Ulmer et al. (2002) Complete kill? √ √ √ √ √ √ × √ √ √ √ >7-log >6-log √ √ >6-log √ √ √ √ √ √ Ref. [42] [43] [34] [46] [87] [52] [35] [85] [56] [81] [28] [90] [30] [31] [65] [61] Leuconostoc dextranicum Listeria innocua Listeria monocytogenes Lin et al. (1993) Dillow et al. (1999) Wei et al. (1991) Lin et al. (1994) Erkmen (2000) Erkmen (2001) [53] [4] [17] [24] [25] [27] Staphylococcus aureus Kamihira et al. (1987) Haas et al. (1989) Erkmen (1997) Dillow et al. (1999) >5-log √ √ √ [15] [28] [20] [4] Staphylococcus saprophyticus Haas et al. (1989) × [28] Table 4 Gram-negative vegetative bacteria species treated with dense-phase carbon dioxide Species Researcher (year) Complete kill? Ref. E. coli Fraser (1951) Kamihira et al. (1987) Haas et al. (1989) Isenschmid et al. (1992) Smelt and Rijke (1992) Ballestra et al. (1996) Shimoda et al. (1998) Debs-Louka et al. (1999) Dillow et al. (1999) Erkmen (2001) Erkmen (2001) Karaman and Erkmen (2001) Schmidt et al. (2005) × >6-log × × >8-log >5-log √ [12] [15] [28] [23] [90] [55] [81] [35] [4] [89] [60] [63] [62] Gluconobacter oxydans Legionella dunnifii Proteus vulgaris Pseudomonas aeruginosa Haas et al. (1989) Dillow et al. (1999) Dillow et al. (1999) Dillow et al. (1999) Spilimbergo et al. (2002) Salmonella salford Salmonella senftenberg Salmonella typhimurium Dillow et al. (1999) Haas et al. (1989) Wei et al. (1991) Erkmen (2000) Erkmen (2001) Serratia marcescens Yersinia enterocolitica Elvassore et al. (2000) Erkmen (2001) × √ √ √ √ √ √ √ √ √ √ √ × √ √ √ √ √ [28] [4] [4] [4] [46] [4] [28] [17] [21] [29] [34] [86] 358 J. Zhang et al. / J. of Supercritical Fluids 38 (2006) 354–372 ever, gram-negative cells have complex cell wall structures but much thinner peptidoglycan layers (only 1–2 layers thick, about 10% of the cell wall) [32,33]. Therefore, the gram-positive cells are stronger, less likely to be broken mechanically, and are less permeable than the gram-negative cells [34]. Both gram-positive and -negative bacteria have been subjected to CO2 treatment (Tables 3 and 4). Generally, gram-positive bacteria have been more difficult to deactivate than gram-negative bacteria. In a semi-continuous apparatus, 107 colony forming unit (cfu) per ml of gram-positive B. subtilis in phosphate buffered saline (PBS) was completely deactivated after 2.5 min exposure under 7.4 MPa CO2 at 38 ◦ C, while the same concentration of gram-negative S. marcescens was deactivated at 0 min (the time when the apparatus is pressurized to the desired pressure) under the same conditions [34]. The same behavior was seen by Dillow et al. [4] and Debs-Louka et al. [35] Electron microscope images provide more information regarding the relationship between the robustness of gram-positive cell walls and reduced cell wall damage. With CO2 treatment at 25 ◦ C, 20.5 MPa for 1 h, SEM images of E. coli (gram-negative), showed partially damaged cell walls and some ruptured cells, while the images of S. aureus (gram-positive) bacterium, showed no cell wall damage [4]. However, gram-negative bacteria are not always more susceptible to high-pressure CO2 than gram-positive bacteria. Dillow et al. examined two gram-positive (L. innocua, S. aureus) and five gram-negative bacteria species (S. salford, P. aeruginosa, E. coli, P. vulgaris, L. dunnifii). Generally, the gram-positive species showed resistance higher, or at least equal to, the gram-negative species. But S. salford, a gram-negative bacterium, showed only 3-log reduction, while the two gram-positive bacteria, L. innocua and S. aureus were reduced by 9- and 7-log, respectively, at 34 ◦ C, 20.5 MPa, 0.6 h, with six pressure cycles [4]. Although there is a difference in sensitivity between the grampositive and -negative vegetative bacteria, both are susceptible to high-pressure CO2 treatment. Twenty out of twenty-two tested vegetative species were completely deactivated at some combination of temperature, pressure, time, etc. (Tables 3 and 4). The species that were not completely deactivated, Salmonella senftenberg and Staphylococcus saprophyticus, have been studied by only one group [28], and the pressures used were only 6.2 and 5.5 MPa. It is possible that the pressures in these studies were too low to give complete deactivation. In conclusion, high-pressure CO2 treatment of vegetative bacteria has been largely successful, with over 90% of the tested species completely deactivated. However, success with vegetative cells does not guarantee that high-pressure CO2 treatment can be used for sterilization. Sterilization is defined as deactivation of ALL living microorganisms including the most resistant form of bacteria — endospores [36]. At least 106 cfu/ml spores must be completely deactivated in order to claim sterilization [4]. 4.3. Treatment of endospores A spore (or endospore) is the highly resistant dormant form of various bacilli and clostridia. Sporulation of vegetative cells occurs under harsh environments such as poor nutrition. Spores Fig. 2. Major structures of a B. subtilis spore (Driks [37]). are highly resistant to heat, UV radiation, free radicals, and chemicals because of their unique structures (Fig. 2) [32,37]. Compared to a vegetative cell which contains on the order of 80–90% water, the spore core is highly dehydrated (only 10–25% water content), making it very resistant to heat and chemicals [32]. The Ca2+ dipicolinic acid complex and small acid-soluble protein (SASP), which bind to DNA, increase spore resistance to heat, desiccation, and UV radiation [32,38,82]. The outside of the spore core is a thick, loosely cross-linked peptidoglycan layer called the spore cortex, which prevents hydration of the spore core [32]. The outmost structure is the multilayered spore coat, which is a permeability barrier to chemicals such as chloroform and lysozyme [37]. Because spores are highly resistant to heat, chemicals and radiation, extreme temperatures (121 ◦ C steam), UV radiation, or highly oxidative chemicals, e.g. ethylene oxide, are used for sterilization. Spore survivability is the standard assay to test sterilization equipment [36]. The most frequently used model organisms are G. stearothermophilus, which is used to test steam and hydrogen peroxide sterilizers, B. atrophaeus, which is used to test dry heat and ethylene oxide sterilizers and B. pumilus, which is used to test radiation sterilizers. Spores have not been studied extensively in the presence of high-pressure CO2 . Only eight species have been investigated in 14 publications (Table 5). Most of the studies on spores report only experimental data and do not address mechanistic questions [4,15,39,40–43]. Only Enomoto et al. [18], Ballestra et al. [44], Watanabe et al. [45], and Spilimbergo et al. [46,47] discuss possible deactivation mechanisms. Spores are highly resistant to high-pressure CO2 treatment. Vegetative G. stearothermophilus cells were reduced by more than 6-log after 1.5-h exposure to CO2 at 2.75 MPa and 25 ◦ C [42]. However, even with 2-h exposure to pure CO2 at 20 MPa and 35 ◦ C, 80% of G. stearothermophilus spores remained viable. Even with the addition of ethanol or acetic acid, less than 60% of the G. stearothermophilus spores were deactivated [15]. J. Zhang et al. / J. of Supercritical Fluids 38 (2006) 354–372 359 Table 5 Bacterial spores treated with dense-phase carbon dioxide Species Researcher (year) Bacillus cereus Dillow et al. (1999) Ishikawa et al. (1997) Watanabe et al. (2003) × [4] [41] [45] Bacillus coagulans Ishikawa et al. (1997) Watanabe et al. (2003) NA × [41] [45] Bacillus licheniformis Bacillus megaterium Watanabe et al. (2003) Enomoto et al. (1997) Enomoto et al. (1997) Ishikawa et al. (1997) × >6-log √ [45] [18] [39] [41] Bacillus polymyxa Geobacillus stearothermophilus Bacillus subtilis Clostridium sporogenes Ishikawa et al. (1997) Kamihira (1987) Roskey and Sikes (1994) Sikes and Martin (1994) Watanabe et al. (2003) Kamihira et al. (1987) Hata et al. (1996) Ishikawa et al. (1997) Ballestra and Cuq (1998) Spilimbergo et al. (2002) Spilimbergo et al. (2003) Parton et al. (2003) Watanabe et al. (2003) Haas et al. (1989) Complete kill? √ √ NA √ 7-log × × >7-log √ √ [15] [28] [23] [51] [16] [56] [80] [40] [18] [57] [81] [35] [34] [54] Torulopsis versatilis Zygosaccharomyces rouxii Candida utilis Shimoda et al. (1998) Shimoda et al. (1998) Isenschmid et al. (1992) Isenschmid et al. (1995) Kluyveromyces fragilis Fungal spores Byssochlamys fulva ascospores Aspergillus niger conidia Penicillium roqueforti spores NA √ √ >8-log √ × √ >4-log √ √ × NA [81] [81] [23] [80] Isenschmid et al. (1992) Isenschmid et al. (1995) × NA [23] [80] Ballestra and Cuq (1998) Kamihira et al. (1987) Ballestra and Cuq (1998) Shimoda et al. (2002) NA >6-log NA >5-log [44] [15] [44] [64] Haas et al. (1989) >6-log [28] 360 J. Zhang et al. / J. of Supercritical Fluids 38 (2006) 354–372 To achieve greater deactivation, several approaches have been employed, including increasing treatment time, raising temperature, and using pressure cycling. Enomoto et al. achieved approximately 7-log reduction of B. megaterium spore with a 50-h treatment at 7.8 MPa and 60 ◦ C [39]. Clearly, a 50-h treatment would be problematic for a practical sterilization process. Spilimbergo et al. reported that only a 0.9-log reduction of B. subtilis spore can be achieved with a treatment at 12 MPa, 54 ◦ C, for 24 h [46]. However, at 75 ◦ C and 7 MPa, greater than 7-log reduction of B. subtilis spores was observed after a 2-h treatment [47]. Ballestra et al. observed a biocidal effect of approximately 3.5-log reduction at 5 MPa for 1 h at temperatures at 80 ◦ C [44]. Hata et al. reported a 6-log reduction of B. subtilis spore with treatment at 70 ◦ C, 20 MPa for 10 h. By increasing temperature and/or the pressure, less time is required to achieve a 6-log reduction. Only 2 h were needed at 90 ◦ C, 6 MPa to achieve a 6-log reduction of B. subtilis spores [40]. The lowest temperature with significant deactivation of B. subtilis spores was reported by Ishikawa et al. [41]. They accomplished a 6-log reduction of B. subtilis spores at 55 ◦ C, 30 MPa for 60 min using their micro-bubble method, in which numerous CO2 micro-bubbles were formed by feeding CO2 through a stainless steel filter (10 ␮m pore size) from the bottom of a pressure chamber. Because these experiments were conducted at moderate to high temperatures, these authors concluded that only with a combination of high-pressure CO2 and at least mild heat can spores be deactivated [46]. Long treatment time and high temperatures are two potential problems of the CO2 sterilization technique. Even though a high degree of deactivation of spores has been realized, this usually requires more than 10 h, which is not competitive with the average time of 10–15 min for steam sterilization. On the other hand, ethylene oxide processes require a 15-h cycle [3]. Additionally, the high temperatures used (55–90 ◦ C) can easily damage heat-sensitive materials. Pressure cycling is a promising method to enhance deactivation while lowering the temperature and time requirements. With pressure cycling of 30 cycles/h, P = 8 MPa, at 36 ◦ C for 30 min, a 3.5-log reduction of B. subtilis spores was achieved. Without pressure cycling, a treatment at 36 ◦ C, 7.5 MPa for 24 h only resulted in 0.5-log reduction [46]. 4.4. Treatment of fungi A few investigators have studied the effect of high-pressure CO2 on fungi and fungal spores. These microorganisms are not the focus of this review, but for completeness, Table 6 shows the references for these studies. 5. Experimental parameters Bacterial cells and spores are complex chemical systems composed of various kinds of organic and inorganic components. When one also considers the variety of growth media, high-phase CO2 processing becomes a very complicated process indeed. Many factors have been studied, including temperature, pressure, depressurization rate, pressure cycling, treatment time, cell concentration, cell growth phase, agitation, media, and entrainer. The influence of these factors will be reviewed below. 5.1. Effects of temperature, pressure, and state Temperature and pressure are the most important factors affecting growth of microorganisms. Each microorganism has a species-specific maximum temperature. Above that temperature, proteins denature, cytoplasmic membranes collapse, and cells lyses and are deactivated [32]. A wide range of temperatures has been employed for high-pressure CO2 treatment, from 0 ◦ C [16] to 100 ◦ C [42,43]. Bacteria are more resistant to pressure than to temperature. A hydrostatic pressure between 100 and 1000 MPa is required to deactivate bacteria [48]. High hydrostatic pressure processes have been reviewed by the Institute of Food Technologists [49] and Cheftel [48]. However, if high-pressure CO2 is used, the pressure requirement can be lowered below 20 MPa (Table 7). The highest pressure reported is only 33 MPa [50]. Generally, deactivation is more pronounced with increasing temperature [21,51]. It is believed that higher temperature enhances deactivation by (a) increasing the fluidity of cell membranes, making them easier to penetrate, and (b) increasing the diffusivity of CO2 [22]. Therefore, higher temperatures reduce the duration of the first stage of deactivation [21], which is thought to be diffusion-controlled (see Section 5.6 for a detailed discussing of two-stage kinetics). Higher temperatures also increase the rate in the second stage [52]. However, higher temperatures may reduce the ability of CO2 to extract low-volatility materials and decrease CO2 solubility in aqueous media [53]. Hong and Pyun reported that deactivation of L. plantarum at 30 ◦ C, 7 MPa was better than that at 40 ◦ C, 7 MPa [31]. This probably is the result of higher density at 30 ◦ C (0.27 g/ml) than that at 40 ◦ C (0.20 g/ml); higher CO2 solubility in the media at 30 ◦ C than that at 40 ◦ C. High pressure facilitates solubilization in water and penetration through cell walls, and increases density and therefore extraction power [24]. All these factors are thought to intensify the deactivation process. Experiments by Debs-Louka et al. showed a pressure threshold below which no deactivation was observed. This pressure threshold varies with bacterial species [35]. The D-value (the time needed to achieve 1-log reduction) of S. cerevisiae showed a steep decrease with increase in pressure from 4 to 10 MPa [54]. The duration of the earlier stage and the inactivation rate of the second stage have been found to be extremely sensitive to pressure [52,55]. Depending on temperature and pressure, CO2 exists in the gas, liquid or supercritical fluid state. Physical properties such as density, diffusivity, solubility in aqueous solution, and extraction power vary dramatically around the critical point [9]. Unfortunately, even though several studies have covered two or three states (Table 7), only a few authors discussed the physical state of CO2 and its possible effect on sterilization [15,53,56]. The supercritical state is characterized by gas-like diffusivity and liquid-like density. The gas-like diffusivity allows supercritical CO2 to quickly diffuse through complex matrices; and the liquid-like density confers high extraction power [9]. Because J. Zhang et al. / J. of Supercritical Fluids 38 (2006) 354–372 361 Table 7 Summary of the high-pressure CO2 experimental conditions of published papers Authors (year) Microorganisms T (◦ C) P (MPa) State t (h) DP PC Media Ref. Freser (1951) Kamihira et al. (1987) Taniguchi et al. (1987) Hass et al. (1989) G− G+ , G− , F, FS 37–38 20–35 35 ∼22–80 3.4 4–20 20 0.3–6.2 G G, L, SC SC G, L 1/20 0–2 2 0.5–168 Fast and slow 20 min NA NA Y N N N [12] [15] [14] [28] Arreola et al. (1991) Wei et al. (1991) G+ , G− 35–60 35 8.3–33.1 5.5–13.6 SC G, SC 1/4–1 1/4–2 Fast NA N N Isenschmid et al. (1992) Lin et al. (1992) Smelt and Rijke (1992) Lin et al. (1993) Lin et al. (1994) Nakamura et al. (1994) Roskey and Sikes (1994) G− , F F G+ , G− G+ G+ F G+ , S 27.5–33 25–35 5–40 25–45 35–45 0–40 3–100 1–15 6.9–20.7 15 6.9–20.7 7.0–21.1 1–4 0.3–6.9 G, L, SC G, L, SC L, SC G, L, SC G, SC G, L G, L 1/12 1/30–1 1/4–1.0 1/120–2/3 0.01–1.0 1/2–5 1–96 NA Fast NA Fast Fast 0.5 MPa/s Fast N Y N Y Y N N Isenschmid et al. (1995) Ishikawa et al. (1995) Sikes and Martin (1994) Ballestra et al. (1996) Hata et al. (1996) Enomoto et al. (1997) F G+ , F G+ , S G− S, F S, F 8–43 25–35 3–100 25–45 35–90 20–60 0 to ∼10 4–25 0.3–7.2 1.2–5.0 4–20 1.0–6.0 G, L, SC G, L, SC G, L, SC G G, SC G 1/12 1/12–1/2 1–96 ∼0.03–1 1/2–30 0–24 N N N N N Y Enomoto et al. (1997) Erkmen (1997) Hong et al. (1997) Ishikawa et al. (1997) Kumagai et al. (1997) Ballestra and Cuq (1998) Shimoda et al. (1998) Debs-Louka et al. (1999) Dillow et al. (1999) S G+ G+ S F S, FS G+ , G− , F G+ , G− , F G+ , G− , S 20–60 25 20–50 31–60 40 50–90 35 Room T 25–60 1.9–9.7 6.0–14.6 2.0–7.0 30 4–15 5 6–30 1.5–5.5 14.0–20.5 G, L G, L G SC G, SC G G, L, SC G L, SC 0–48 0–5 1–4 0–1.33 0–5 1/4–1.0 1/4–1/2 0.31–6.19 0–4 7.3 MPa/min 5 min Fast NA 8 MPa/h 4.8 MPa/min, 0.033 MPa/min 60 s 60 s NA 5 min 8 MPa/h NA Fast and slow 0.4 s NA Broth Wet and dry Wet and dry Juice, broth, H2 O, PS, foods, Wet and dry Juice Juice, broth, H2 O, meat, egg Broth Broth Agar, buffer Broth Broth, milk Wet and dry Buffer, broth, H2 O Broth PS Buffer, broth Buffer H2 O, peptone H2 O Hong et al. (1999) Hong and Pyun (1999) G+ G+ 25–45 20–40 6.9–13.8 5.0–8.0 G, L, SC G, L, SC 0–1 0–4 NA 2 min N N Elvassore et al. (2000) Erkmen (2000) G+ , G− , F G+ 38–40 25–45 5.5–8.0 1.5–6.0 G, SL G 0–0.5 0–24 NA