Substantia. An International Journal of the History of Chemistry 4(2) Suppl.: 57-67, 2020 Firenze University Press www.fupress.com/substantia ISSN 2532-3997 (online) | DOI: 10.36253/Substantia-1030 Citation: A. Garrido Sanchis, B.W. Ninham (2020) Very High Efficiency of E. coli Inactivation by Body Temperature CO2 Bubbles: in Pursuit of Mecha- nism. Substantia 4(2) Suppl.: 57-67. doi: 10.36253/Substantia-1030 Copyright: © 2020 A. Garrido Sanchis, B.W. Ninham. This is an open access, peer-reviewed article published by Firenze University Press (http://www. fupress.com/substantia) and distributed under the terms of the Creative Com- mons Attribution License, which per- mits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All rel- evant data are within the paper and its Supporting Information files. Competing Interests: The Author(s) declare(s) no conflict of interest. Very High Efficiency of E. coli Inactivation by Body Temperature CO2 Bubbles: in Pursuit of Mechanism Adrian Garrido Sanchis2,*, Barry W. Ninham1 1 Department of Applied Mathematics, Research School of Physical Sciences, The Austral- ian National University, Canberra, Australia 2 School of Sciences, University of New South Wales, Canberra, Australia *Corresponding author: a.garridosanchis@adfa.edu.au Abstract. A CO2 bubble column (CBC) has been developed as a body-tempera- ture lab-scale water sterilization process for the inactivation of pathogens. Both CO2 and combustion gas bubbles inactivated Escherichia coli C-3000 (ATCC15597) with extraordinary efficiency in solutions with low alkalinity. The mechanisms of inactiva- tion were not known. To characterise the phenomena a new first-order kinetic equa- tion that correlates E. coli inactivation rates with a total alkalinity of the solutions has been developed as a first step towards understanding. This leads us to propose a new mechanism of inactivation. Keywords: E. coli, water reuse, carbon dioxide, combustion gas, alkalinity. 1. INTRODUCTION Water-borne pathogens remain one of the leading risks to public health worldwide. The World Health Organisation (WHO) estimated that over 30 million cases of diseases and millions of deaths could be caused by patho- gen-infected water sources, globally, each year.1 The inactivation of microorganisms like bacteria and viruses in aqueous solutions is a perpetual challenge. For example, pathogenic bacteria present in water used in food or pharmaceutical manufacture can cause product contamination. In a bigger scenario wastewater from agricultural or indus- trial uses or water obtained from environmental flows, contains pathogenic microorganisms that need to be inactivated before the water is used for industrial or agricultural purposes or drinking water. This is an existential problem for humanity. Present methods for inactivating viruses, bacteria, and other microorgan- isms include heat, irradiation (e.g. ultraviolet treatment), treatment with chem- icals (e.g. ozone), high-pressure treatment, and filtration (e.g. membrane filtra- tion). Many of these methods, especially, heat treatment, are energy-intensive. More energy-efficient treatment technologies are desperately needed. 58 Adrian Garrido Sanchis, Barry W. Ninham High density and high-pressure CO2, known as dense phase carbon dioxide (DPCD), is being used as a non-thermal disinfection alternative.2,3 Our process is different. We have previously stud- ied CO2 sterilization effects at atmospheric pressure in a bubble column on E. coli and viruses. The process has been implemented commercially. We found that for both pathogenic groups the sterilization effects increase with the temperature.4 It is of much interest that inac- tivation becomes optimal above 38 degrees, body range temperature. Carbon dioxide, being a greenhouse gas, is assigned some responsibility for global warming. But it is also lit- erally vital for plant growth. Recently it has been shown likely to be Nature’s way of controlling viral infection in the human body, following from these CO2 antiviral properties a recent study has identified the use of body temperate humidified CO2 as a possible early therapy to mitigate infection in COVID-19 patients.5, 6 Many alter- natives to reduce CO2 emissions have been proposed, apart from planting trees and controlling bushfires. Geological sequestration in depleted gas and oil reser- voirs, ocean storage, seaweed production, CO2 minerali- zation are just a few. Sterilisation by CO2 has been widely used through- out history. After sodium chloride, natron, sodium bicarbonate is probably the first industrial-scale chemi- cal used by mankind. Natron, in ancient Egypt, was essential for mummification. The still and little under- stood peculiar hydration properties of the bicarbonate ion in contact with air affects both tissue dehydration and sterilization against bacterial degradation. The magic of the bicarbonate ion has remained, with the still unexplained problem of Roman and other cement formation. The foundation of physical chem- istry can arguably be marked by Berthollot’s observa- tion that reactions are temperature dependent: He saw soda-lime, sodium carbonate rocks instead of the cal- cium carbonate on the banks of the Nile River during Napoleon’s scientific expedition to Egypt in 1792. The normal precipitation of calcium carbonate from a sodi- um chloride calcium carbonate solution is reversed at high temperatures of the drying Nile flood in summer. The astute chef and detergent companies know that wet sodium bicarbonate is extremely effective in cleaning. It cuts peptide bonds of proteins bound to saucepans or dishes. But why? 1.1. Nature of the Present Studies The work here presented studies the E. coli inacti- vation effect of CO2 at different alkalinity levels at 38 °C (human body temperature range) and atmospheric pressure. Alkalinity is defined as the ability of water to resist changes in pH when acids or bases are added to the solu- tion or the capacity of water to neutralize acids. The CO2- carbonate system describes which of the four forms of total inorganic carbon (molecular aqueous CO2, Carbon- ate ion, CO3-2, Bicarbonate ion, HCO3-, or Carbonic acid, H2CO3) is the dominant species as a function of pH.7,8 We have analysed our results within the boundaries of the accepted theory of physical chemistry of solutions. This allowed the development of a new first-order kinet- ics equation that correlates E. coli inactivation rates with total alkalinity when bubbling CO2 or combustion gases in a bubble column at atmospheric pressure and body temperature. But while it fits the data it does not account for why the process is efficient. We are then led to such an explanation outside the confines of classical theory. 2. MATERIALS AND METHODS 2.1. Coliform bacteria Escherichia coli is a gram-negative bacterium with a straight cylindrical rod shape of 1.0-2.0 µm size.9 E. coli are harmless, can be found in the intestines of healthy humans and animals but some are pathogenic to the host. As a result of fecal contamination, it can be found in water and soil. Its presence in water is used as an indicator to monitor water quality.10 The strain E. coli C-3000 (ATCC15597) is a biosafety Level-1 organism11 which was selected as a representative model for bacterial contamination in water.12,13 Spinks at al.14 demonstrated that pathogenic bacte- ria are inactivated in a temperature range of 55 to 65 °C. Other studies found that E. coli presents the first signs of thermal inactivation at temperatures over 55 °C, with high inactivation rates at 60 °C.15 An overnight E. coli C-3000 (ATCC 15597) cul- ture was grown in 10 mL broth (see section 2.3 for more details) without agar at 37 °C for 18–20 hours in a Labtech digital incubator, model LIB-030M, while shak- en at 110 rpm by a PSU-10i orbital shaker. 2.2. Experimental solutions Solutions with different alkalinity levels (see Table 1) were prepared and sterilized by autoclaving in an Aescu- lap 420 at 15 psi and 121–124 °C, for 15 minutes.16 A standard secondary-treated synthetic sewage medium (Organisation for Economic Co-operation and 59Very High Efficiency of E. coli Inactivation by Body Temperature CO2 Bubbles Development (OECD) medium) was prepared accord- ing to water-quality guidelines and standards.17,18 The official Journal of the European Community for Sec- ondary-treated Water Quality has the following require- ments for discharges from urban wastewater-treatment plants: 125 mg/L of COD; 2 mg/L of total phospho- rus; and 15 mg/L of total nitrogen.19 This solution was designed to meet the European standards by using the following ingredients in a liter of boiled tap water or Milli-Q water (depending on the experiment): 120 mg of peptone; 90 mg of meat extract (replaced here by Bovril® according to recommendations in Biology of Wastewater Treatment20); 30 mg of urea; 13 mg of dipo- tassium hydrogen phosphate; 7 mg of sodium chloride; 2 mg of calcium chloride dehydrate; and 2 mg of mag- nesium sulfate heptahydrate. Two NaCl solutions with concentrations of 0.001 M and 0.17 M NaCl (NaCl ≥ 99% purity, obtained from Sigma-Aldrich) in Milli-Q water, a 0.01 M CaCl2 solu- tion (CaCl2.≥ 99% purity, obtained from Sigma-Aldrich) in Milli-Q water and tap water. These concentrations were chosen to coincide roughly with the physiological concentration of humans and to typical environmental conditions faced by the bacterium. To ensure the full elimination of chlorine from tap water it was previously boiled for 30 minutes, there- fore E. coli (C3000) was not affected by the disinfecting action of chlorine. 2.3. E. coli growth media The plate count method21 is commonly used for the detection of E. coli in, both wastewater and marine water. The degree of contamination in water is assessed based on the ability of bacteria to propagate forming col- onies in a layer of agar.22 For each bacterial-growth experiment, two broths were prepared (A and B). Broth A was prepared with 1 g of yeast extract, 6 g of NaCl, 13 g of tryptone, and 1,000 mL of Milli-Q water. This medium was aseptically distributed into two vessels containing 1.41% agar and no agar, respectively; the agar used in the experiments was molecular-biology grade from Sigma-Aldrich. The medium was heated to boiling to dissolve the agar and sterilized by autoclaving for 15 minutes in an Aesculap 420 at 15 psi and 121–124 °C. Broth B was used to improve the viability of the bacteria. It was prepared by adding 0.010 g of thiamine and 1 g of glucose to 50 mL of Milli-Q water and filtered through a 0.22 µm filter for its sterilization. Once cooled to 50 °C, it was added aseptically to broth A in a propor- tion of 1:19. The resulting 1.34% agar medium was poured into 100mm × 15mm Petri dishes and dried above a Bunsen burner to maintain local environmental sterility until the agar was not too dry nor too moist.23 2.4. CO2 Bubble column (CBC) process The CBC process used two different gases: CO2 and combustion gases. A flow of 24 L/min of CO2 was pumped through the electrical heater that maintained the gas temperature just above the sinter surface at 38°C for the CO2 experiments. When using combustion gases, the exhaust pipe of an LPG gas generator (Greenpower 2.5 kW) was attached to an isolated metal pipe with a valve that provided an exhaust-gas flow rate of 27 L/min through the bubble column reaching a temperature of 58 °C just above the sinter surface. The base of the bubble column evaporator was fitted with a glass sinter (type 2) of 135 mm diameter and pore size 40–100 µm. Once the experimental solutions (0.17 M NaCl, 0.01 M CaCl2, 0.001 M NaCl, secondary treated synthetic sewage made with Milli-Q water and tap water) were poured into the column, the temperature of the solution was measured with a thermocouple in the center of the column. Hot bubbles of both gases passed through the sinter into the 300 ml solution, inactivating E. coli. The sterilization effects of other gases like air, argon, and nitrogen on E. coli have been discussed24, they are not significant compared with CO2 bubbles, especially at low temperatures. 2.5. Experimental setup and procedure The CBC treated 0.17 M NaCl, 0.01 M CaCl2, 0.001 M NaCl, secondary treated synthetic sewage made with Milli-Q water, and tap water containing E. coli C-3000 (ATCC15597) at inlet CO2 temperature of 38 °C, and combustion exhaust gas from an LPG generator at 60 °C. The plate count method25,26,27 was used to assess the viability of E. coli. Samples were collected from 10–15 mm above the central area of the bubble column. Each sample of 0.1 mL was spotted in triplicate. Alkalinity measures the acid-neutralizing abil- ity of dissolved substances in water and is equivalent to the amount of strong acid required to lower the pH of the solution from an initial value to about 4.2. The total alkalinity of each solution, in mg/L CaCO3, was deter- mined by the amount of 0.02 M HCl added in a titration until a pH electrode (ThermoFisher Scientific) measured an end pH of 4.2. The initial pH of the sample of each of the 6 solutions was less than 8.3 therefore the alkalin- 60 Adrian Garrido Sanchis, Barry W. Ninham ity was calculated with the carbonate ion concentration multiplied by a factor of two since each [CO3-2] ion neu- tralised two protons28, see Table 1. 2.6. Data analysis The linear decay model was used to study the inac- tivation of E. coli over time in CBC.24,29,30 The evaluation of viable E. coli in the samples was performed using the plate count agar (PCA) method.23,26,31 Plate counts num- bers indicate the number of E. coli colonies that grow on the surface of a solid and translucid agar media in a Pet- ry dish. The mean and standard deviation of each tripli- cated sample was obtained using bacteria survival factor, log10(CFU/ CFU0), where CFU0 is the initial number of colony-forming units (CFU) per sample and CFU is the number of E. coli colonies per sample after a set expo- sure time to the CO2 bubbles.32 When analysing the data we assumed that the inac- tivation of E. coli by CO2 in different alkaline solutions follows first-order kinetics. We proposed a model that uses the decimal reduction time (D-value) to measure how the alkalinity of a given solution can influence E. coli inactivation. The D-Value is defined as the time, in minutes, needed to inactivate 90% (i.e. 1-log ) loss of E. coli. D-values were calculated using a linear exponential decay model. log(CFUt)=log(CFU0)– [1] [2] Here, is the number of E. coli colonies at time t, = the initial number of E. coli colonies, D= the decimal reduction time, -(1/D) = the slope of the curve. 3. RESULTS AND DISCUSSION 3.1. CBC process inactivating E. coli at 38 °C and one atmosphere Zhong et. al.33 observed that when bubbling CO2 in different electrolyte solutions, their pH dropped until it reached a steady value in less than 10 minutes; this value was different for each solution. Similar behavior was observed when CO2 gas at 38 °C and combustion gases were bubbled for 10 minutes through six different solutions (0.17 M NaCl, 0.01 M CaCl2, 0.001 M NaCl, secondary treated synthetic sewage made with Milli-Q water and tap water) in the bubble column (see Table 1). The absorption of CO2 into different solutions increases when CO2 gas is bubbled through the sintering area in a bubble column; many bubbles are produced with a large CO2-liquid contact surface that is continually produced. When CO2 dissolves in water 99% stays as the dis- solved molecular gas and less than 1% as carbonic acid (H2CO3 (equations 1 and 2) which reduces the pH of the solution (see Table 1). Carbonic acid dissociates into bicarbonate ion (HCO3-) and carbonate ion (CO32-) (see equations 3 and 4).34 CO2(g)→CO2(aq) [3] CO2(aq)+H2O(l)⇆H2CO3(aq) [4] H2CO3(aq)⇆H++HCO3-(aq) [5] HCO3-(aq)⇆H++CO32-(aq) [6] This is the reason why the pH of every solution dropped (see Table 1) and was different from that of the other solutions.34 When a 24 L/min flow of CO2 passes through a sin- gle glass tube immersed in a beaker containing E. coli in 300 mL of 0.17 M NaCl the big bubbles of 1-3 cm diam- eter do not inactivate the pathogens.4 However, when the same flow of CO2 gas at the same temperature in the same solution is bubbled through a sintering area in a bubble column E. coli inactivation occurs. The bubble column produces many small CO2 bubbles (1-3 mm diameter). The consequent CO2 dissolution rate increases due to the large CO2-liquid contact surface that is constantly produced. Mass transfer from the CO2 to the liquid depends high- ly on the interfacial area (α).4 This increases the amount of CO2 dissolved in the solution and produces a simi- lar sterilization effect to what can be achieved by raising the pressure in dense phase carbon dioxide (DPCD) pro- cesses. The overwhelming advantage of the bubble column delivery is that only atmospheric pressure is required. A possible mechanism previously described4, could explain in part why E. coli (doesn’t work for viruses) can be inactivated by bubbling CO2 at body temperature in a bubble column. It is that the pH decreases when CO2 dissolves in a solution and this acidification of the solu- tion increases the binding of bicarbonate ion to bacte- rial membrane phospholipid head groups. The resulting change in the head group area and membrane structure to an open mesh phase then allows free diffusion of CO2 through the bacterial membranes. The CO2 inside the cell produces an intracellular pH decrease that exceeds the cell’s buffering capacity, resulting in cell inactiva- 61Very High Efficiency of E. coli Inactivation by Body Temperature CO2 Bubbles tion.35,36 Orij et al. observed that Saccharomyces cerevi- siae cell division rate is controlled by intracellular pH changes and low intracellular pH could reduce the cell division.37 This idea has a degree of plausibility and may well contribute to bacterial inactivation. However, it does not explain the inactivation of viruses. When CO2 was bubbled at 38 °C through 0.17M NaCl and 0.001 M NaCl solutions (made with Milli-Q water) in the bubble column evaporator, for both solu- tions, 0.6-log of E. coli inactivation was observed after 10 minutes due to the high CO2 surface area in the solution (see Fig. 1). This occurred even though the two solutions had very different bubble densities. When using second- ary-treated synthetic sewage and 0.01 M CaCl2 (both made with Milli-Q water), a 0.2-log and 0.3 log reduction of E. coli was achieved respectively after 10 minutes (Fig. 1a and Table 1). However, when the same experiment was conducted with boiled tap water and with secondary- treated synthetic sewage (made with boiled tap water), no inactivation was observed after 11 minutes (Fig. 1a). When combustion gas from a generator (with 12.5–14% of CO2)38 at 60°C was used in the bubble col- umn experiments an E. coli inactivation rate of 2.6-log in 0.17 M NaCl solutions (made with Milli-Q water) was observed after 10 minutes.24 However, when using sec- ondary-treated synthetic sewage (made with boiled tap water) no inactivation was detected after the same peri- od (Fig. 1b). This is consistent with the hypothesis that reduced pH buffering capacity aids inactivation. With- in the combustion gas, other products than CO2, H2O, and N2 could be present due to minor components and impurities in the fuel and different fuel/air ratios. These gases are carbon monoxide (CO), hydrogen (H2), sulfur oxide (SO2), and mono-nitrogen oxides (NOx) such as NO and NO2.38 Even with the presence of these gases, no inactivation was observed. The tap water used in this work had a total alkalin- ity of 87.4 mg CaCO3/L (Table 1), and an initial pH of 8 and therefore can act as a buffer solution. After bubbling CO2 the pH was reduced to 4.9 for the pure tap water and 5.4 for the secondary-treated synthetic sewage, with an alkalinity of 94.5 mg CaCO3/L (Table 1). This pH is too high to allow the penetration of the CO2 through the membranes, with a consequent lack of E. coli inactiva- tion when using the CBC process in solutions with high alkalinity. Milli-Q-water-based solutions had a reduced buffer capacity of only 6 mg CaCO3/L (Table 1), reaching low pH values of 4 for NaCl and CaCl2 based solutions. For Milli-Q-water based secondary-treated synthetic sew- age the pH reached 4.4 and the total alkalinity value was 27.7 mg CaCO3/L (Table 1); this solution contains differ- ent salts that can slightly increase its alkalinity, with a consequent reduction in inactivation. 3.2. Effect of alkalinity on E. coli inactivation when bub- bling CO2 and combustion gases Decimal reduction times (D-values) (Table 1), calcu- lated from inactivation values, have been used to under- stand the effect of alkalinity on E. coli inactivation. To do this, we compared E. coli inactivation performance of the CBC when using body temperature CO2 with that of combustion gases for six solutions (0.17 M NaCl, 0.01 M CaCl2, 0.001 M NaCl, secondary treated synthetic sew- age made with Milli-Q water and tap water) with differ- ent alkalinity values, see Fig. 2. The correlation between the log of the D-values and the corresponding total alkalinity is shown in Fig. 2 and Table 1. A D-value is a time needed to inactivate 90% (i.e. 1-log) of the E. coli. To measure the impact of alkalinity on E. coli inactivation when bubbling CO2 we have proposed Figure 1. (a) Comparison of E. coli inactivation with CO2 at 38 °C in a bubble column in 6 different solutions; (b) E. coli inactivation with combustion gas at 60 °C in a bubble column in 2 different solutions. 62 Adrian Garrido Sanchis, Barry W. Ninham the new AL-value. Th e AL-value gives the total alkalinity required to change the D-value by a factor of 10 and refl ects the alkalinity impact on E. coli inactivation. Th e smaller the AL-value, the greater the sensitivity to low alkalinity. Figure 2 compares the impact of alkalinity on E. coli inactivation between CO2 and combustion gases. Above and to the right of the lines E. coli will be inactivated by at least 1-log. From the AL-values obtained with CO2 (AL=40.5) and combustion gases (AL=35), it can be seen that the inactivation for both gases showed a similar alkalinity- dependent behavior (Fig. 2). A reduced AL-value of 35 to 40.5 mg/L of CaCO3 suggests that when bubbling both gases inactivation rates of E. coli will highly depend on the alkalinity of the solution, see Table 1 and Figure 1. For a lka linit y va lues in the range, 5–95 mg CaCO3/L, combustion gases presented higher E. coli inactivation rates than when using pure CO2, with D-values of 16.3 min at 6 mg CaCO3/L for pure CO2 and 4.6 min at 6 mg CaCO3/L for combustion gases. At higher alkalinity values both gases needed longer times to achieve 1 log E. coli inactivation, combustion gases required 1,666 min at 95.4 mg CaCO3/L, pure CO2 need- ed 3,333 min at 95.4 mg CaCO3/L (Fig. 2 and Table 1). For both gases, E. coli inactivation can be increased by reducing the alkalinity of the solution. Solutions with a low buff er capacity, low alkalinity, aft er bubbling CO2, or combustion gases in the bubble column for 10 min- utes signifi cantly reduced their pH as observed in our experiments (Table 1). This reduced pH could affect microbial cell inactivation, since cell membranes stop protons from penetration but also make them more permeable to other substances, like CO2, due to the chemical modifi cation on the phospholipid bilayer of the membranes.35,36 When CO2 penetrates inside of the bacteria dissolves in the intracellular fl uid producing carbonic acid (H2CO3) what triggers an intracellular pH decrease that exceeds the cell’s buff ering capacity, result- ing in cell inactivation.35,36 3.3. Development of an inactivation model based on alka- linity Based on the evidence (see Figure 3) that E. coli inac- tivation in diff erent alkaline solutions when bubbling 38 Table 1. Initial and fi nal pH values aft er bubbling CO2 and combustion gas in 6 diff erent solutions, amount of 0.02 M HCl to reduce the pH of the solutions to 4.2, alkalinity in mg/L of CaCO3 and D-values. Solution pH initial pH fi nal ml of HCl 0.02 M to reach pH 4.2 ml of HCl 0.02 M to reach fi nal pH Total Alkalinity (mg/L of CaCO3) (to reach a pH of 4.2) Equivalent (mg/L of CaCO3) (to reach fi nal pH) D-value (min) Log10 (D-value, min) CO2 gas at 38°C 0.001M NaCl in Milli-Q water 5.78 4 0.9 1.3 6 8.7 16.3 1.21 0.17M NaCl in Milli-Q water 5.86 4 0.9 1.3 6 8.7 16.6 1.22 0.01M CaCl2 in Milli-Q water 5.94 4 1 1.5 6.7 10 34.1 1.53 Tap water 8 4.9 13.1 12 87.4 80.1 2,000 3.30 Secondary-treated synthetic sewage made with tap water 7.63 5.4 14.3 10.5 95.4 70.1 3,333 3.52 Secondary-treated synthetic sewage made with Milli-Q Water 6.83 4.4 4 2.6 26.7 17.3 54.9 1.74 Combustion gas at 60 °C Secondary-treated synthetic sewage made with tap water 7.63 6.5 14.3 5.4 95.4 36 1,666 3.22 0.17M NaCl in Milli-Q water 5.86 4 0.9 1.3 6 8.7 4.6 0.66 Figure 2. Minimum CO2 bubbling times at diff erent alkalinity val- ues to achieve 1-log bacteria inactivation in 6 diff erent alkaline solutions. Above and to the right of the lines the pathogens will be sterilized by 1-log. 63Very High Efficiency of E. coli Inactivation by Body Temperature CO2 Bubbles °C CO2 and combustion gases follows first-order kinetics we proposed a new E. coli inactivation equation that con- siders the alkalinity of different solutions. 3.4 Disinfection Kinetics model The disinfection efficiency of bubbling CO2 or com- bustion gases at low temperatures through a solution with E. coli depends on the total alkalinity of the solu- tion. A new model that includes alkalinity based on the Chick-Watson first-order equation has been developed in this work. E. coli inactivation follows the first-order kinetics, thus the model results in the following equation: =–k CFU [7] where the E. coli population (CFU) varies with process- ing time (t) at a constant rate (k). The New Model based on the decimal reduction time (D-value) and the new AL-value explains how the alkalinity of a given solution can influence E. coli inacti- vation when CO2 or combustion gases, that contain CO2, are bubbled through at atmospheric pressure. The D-val- ue is defined as the time, in minutes, needed to inacti- vate 90% (i.e. 1-log) of E. coli. D-values were calculated using a linear exponential decay model. logS(t)=log [8] where S(t) is the survival fraction after a time (min), is the number of E. coli colonies at time t, = the initial number of E. coli colonies, D= the decimal reduction time, -(1/D) = the slope of the curve (see Figure 3). For this work, we have defined a new parameter, the AL-value, that is the reduction in alkalinity needed to reduce the D-value by 1-log when bubbling CO2 in a bubble column at body temperature (38°C) and atmos- pheric pressure. It measures the impact of alkalinity on E. coli inactivation. Thus: logD1= (A1–A2)+logD2 [9] AL= [10] Where, A1 first alkalinity of the interval, A2 second alkalinity of the interval, and D1 and D2 are the D-values at A1 and A2. Equation 5 is obtained from Fig. 3, where A is alka- linity. logD=0.0253A+1.0087 [11] From quations 2 and 5. [12] Where, CFUt is the number of E. coli at time t, CFU0= the initial number, A= is the alkalinity of the solution in mg/L of CaCO3 (see Figure 3). 4. POSSIBLE MECHANISMS OF STERILISATION 4.1. Negotiating the Quagmire: Nanobubbles We began this and previous works4,24 with the remarkable discovery that CO2 bubbles at body temper- atures are extremely efficient sterilising agents for both bacteria and viruses. The technology is being imple- mented at industrial scales. Potential applications are legion. The use of recycled water in cities has been pro- hibited in large part because viruses could not be eas- ily removed from drinking water. A recent discovery in physiology has revealed the structure and function of the ubiquitous endothelial surface layer (ELS), unknown for 150 years.39 It is a micron thick foam of carbon dioxide nano- bubbles. CO2 is being produced by metabolism. and passes through a molecular frit that is the polymeric gly- cocalyx to form nanobubbles. This mimics our bubble column apparatus and vice versa. The ESL foam repels red cells and protects organs from invasion by bacteria, neutrophils, low-density lipoproteins, and a legion of Figure 3. Minimum CO2 and combustion gas bubbling times at dif- ferent alkalinity values to achieve 1-log bacteria inactivation in 6 different alkaline solutions. Above and to the right of the lines the pathogens will be sterilized by 1-log. 64 Adrian Garrido Sanchis, Barry W. Ninham other invaders, good and bad. Viruses like COVID-19 have to run the gauntlet of this layer of nanobubbles on their way out to the adjoining bloodstream and exit with CO2 nanobubbles via the lungs. The issue of CO2 sterilisation takes on a wider import as Nature’s gift to protect us from the disease. The question of what fac- tors combine to account for the mechanism of sterilisa- tion then takes on a decided imperative. The results of the experiments we have been able to subsume in the single first-order kinetic equation of Section 3.3. We now have to see if we can extract any sense and hints at mechanism from this. In attempting to do so we have negotiated our way through a swamp of misinformation. By this, we mean the classical foundations of physical chemistry that inform our intuition suffer from seri- ous sins of omission and of commission. 40 These are due to the omission of specific ion (Hofmeister) effects, of dissolved gas: and the organization and reactivity of nanobubbles. Their universal presence has been a hidden variable that controls reactivity in solution. They are the energy source via free radicals that drive in polymeriza- tion41, and enzymatic reactions.42,43 Their presence has been revealed and studied extensively by laser spectros- copy over many years.44-46 Nanobubbles even appear on mechanical shaking and contain highly reactive hydro- gen peroxide and of course on sonication.47-49 4.2. Available Evidence With these complications of classical theory in mind we marshall what further evidence we can that could illuminate the issue of mechanism. Other gases: argon, air, and nitrogen have no dele- terious on pathogens at low temperatures. At very high temperatures they do explore in a succeeding paper.24 Our problem is peculiar to CO2. 4.3. Sodium bicarbonate. The astute chef and detergent companies know that the removal of food proteins adhering to utensils is not easily accomplished. It is by using reasonably warm con- centrated sodium bicarbonate solution. This involves active nanobubbles nucleated at the surface that cut pep- tide bonds. It is also aided by Berthelot’s 1795 observa- tions on the Nile of precipitation of sodium carbonate rather than calcium carbonate. The ion exchange at and above 50 degrees removes calcium and weakens adhesive bonds. This occurs with the destruction of proteins and their adhesion in the protective coats of both bacteria and viruses. With bacteria the lipids of the exterior membranes of both gram-positive and gram-negative types fall apart above 50-60 degrees because of the same change in head group hydration with temperature – apart from the obvious change in hydrocarbon chain packing. 4.4. pH changes with salt and salt concentration We have already remarked on apparently peculiar specific differences in pH with different salts with bub- bling CO2 through the solution. This artifact occurs for solutions with CO2 and those without. They are because the classical theory of pH and buffers assumes that elec- trostatic forces alone are operating. Specific Hofmeister effects arise when the theory is corrected to include dis- persion force acting ions, anion cations, and buffer ions. Further complications of buffers show up explicitly in42. But also listed below of course all classical measure- ments ignore dissolved gas that also affects matters like the interpretation of surface potentials (pH). Reversal of the Hofmeister series for pH is seen with buffers nomi- nally at the same pH, e.g., phosphate vs cacodylate, nom- inally pH 7.50-52 These are serious complications around which we have to tread carefully. See e.g. for an explicit illustra- tion of obstacles to interpretation caused by the use of classical theory.53 This reports studies of growth rates of Staphylococcus aureus and Pseudomonas aeruginosa as a function of salt concentration for a large range of anions spanning the Hofmeister series. The variations are enor- mous and not explainable in the usual framework of col- loid science.53 4.5. Finally a Mechanism of Sterilisation So armed for pitfalls of theory, we can tread careful- ly towards an understanding of what is going on. Essentially this: Nanobubbles are formed through the turbulent passage of CO2 through the bubble column frit. They are stabilised by the adsorption of ions and proteins from the bubble column solution. The nanobub- bles will have a dynamic highly reactive surface contain- ing bicarbonate and carbonate ions in amounts depend- ing on alkalinity that determines the proportion of each. Carboxylates form due to surface reactions with solutions depending on solution constituents. If in high enough concentration these will react with peptides of bacterial (and viral) coats, resulting in destruction. This is consistent with the astute chef and detergent compa- nies cleaning recipe. 65Very High Efficiency of E. coli Inactivation by Body Temperature CO2 Bubbles 5. CONCLUSIONS Pure CO2 gas or combustion gases can be used in the bubble column to effectively inactivate E. coli C-3000 (ATCC15597) in the water at atmospheric pressure and ambient temperatures. The efficiency of the process appears to depend on the alkalinity of the solution. CO2 bubbles will have a dynamic highly reactive surface containing bicarbonate and carbonate ions in amounts depending on the alkalinity of the solution that determines the proportion of each. Carboxylates form due to surface reactions with solutions on solution con- stituents. If in high enough concentration, especially in low alkalinity solutions, this will react with peptides of bacterial (and viral) coats, resulting in their inactivation. We have developed a new first-order kinetics equa- tion that predicts E. coli inactivation rates from solutions with different alkalinity when bubbling CO2 or combus- tion gases in a bubble column at atmospheric pressure. 6. ADDENDUM Since this work was written we discovered an excel- lent paper from 1948 dealing with the Surface Inac- tivation of Bacterial Viruses and of Proteins by Mark H.Adams.54 It explores the effects of shaking with differ- ent gases but not CO2 on viruses and proteins. It even compares with the effect of bubbling through a frit. This and references contained therein go back to the begin- nings of genuine molecular biophysics and predate the molecular biology DNA revolution. It would be remiss of us not to refer the reader to the beginnings of quan- tification of the physical effects of gas and nanobubbles, long-forgotten and missing from contemporary physical chemistry. 7. CONTRIBUTIONS Adrian Garrido Sanchis (principal author) carried out the lab experiments, performed analysis on all samples, interpreted data, and lead the writing of the manuscript. Barry Ninham acted the part of devil advocate and critic and led his disciple back to the paths of correctness. 8. REFERENCES 1. WHO, Technical guidance on water-related disease surveillance 2011, World Health Organization. 2. J.F. Rusling, Environmental Electrochemistry: Funda- mentals and Applications in Pollution Abatement By Krishnan Rajeshwar (University of Texas at Arling- ton) and Jorge G. Ibanez (Universidad IberoAmeri- cana). Academic Press: San Diego. 1996. 776 pp. ISBN 0-12-576260-7, J Am Chem Soc, 1998, 120(45), 11837-11837. 3. G.F.a.M.O. Balaban, Introduction to Dense Phase Carbon Dioxide Technology, in Dense Phase Carbon Dioxide, B.P. Professional, Editor. 2012. p. 1-4. 4. A. Garrido Sanchis, R. Pashley, B.W. Ninham, Virus and bacteria inactivation by CO2 bubbles in solution, npj Clean Water, 2019, 2(1), 5. 5. B.P. Reines, B.W. Ninham, Structure and function of the endothelial surface layer: unraveling the nano- architecture of biological surfaces, Quarterly reviews of biophysics, 2019, 52, e13. 6. A. El-Betany, E. Behiry, M. Gumbleton, K. Harding, Humidified Warmed CO2 Treatment Therapy Strate- gies Can Save Lives with Mitigation and Suppression of SARS-CoV-2 Infection: An Evidence Review. 2020. 7. G.W. VanLoon, S.J. Duffy, Environmental chemistry : a global perspective, 2005. 8. C. Baird, M. Cann, Environmental chemistry. 2012, New York, NY: W.H. Freeman and Company. 9. G. Reshes, S. Vanounou, I. Fishov, M. Feingold, Cell shape dynamics in Escherichia coli, Biophys J, 2008, 94(1), 251-264. 10. R.A. Welch, The Genus Escherichia, in The Prokary- otes: Volume 6: Proteobacteria: Gamma Subclass, M. Dworkin, et al., Editors. 2006, Springer New York: New York, NY. p. 60-71. 11. ATCC, Product Sheet Escherichia coli (ATCC 15597), ATCC, Editor. 2015. 12. X. Yang, Introduction, in A Study on Antimicrobial Effects of Nanosilver for Drinking Water Disinfection. 2017, Springer Singapore: Singapore. p. 1-12. 13. I. Gaska, O. Bilenko, S. Smetona, Y. Bilenko, R. Gas- ka, M. Shur, Deep UV LEDs for Public Health Appli- cations, Int J High Speed Electron, 2014, 23(03n04), 1450018. 14. A.T. Spinks, R.H. Dunstan, T. Harrison, P. Coombes, G. Kuczera, Thermal inactivation of water-borne pathogenic and indicator bacteria at sub-boiling tem- peratures, Water Res, 2006, 40(6), 1326-1332. 15. McGuigan, Joyce, Conroy, Gillespie, M. Elmore, Solar disinfection of drinking water contained in transparent plastic bottles : characterizing the bacterial inactivation process, J Appl Microbiol, 1998, 84(6), 1138-1148. 16. I. 10705-1, Water quality - Detection and enumeration of bacteriphages- Part 1, in ISO 10705-1. 1995, Inter- national Organization for Standardization: ISO. 17. I. 11733, Water quality - Determination of the elimi- nation and biodegradability of organic compounds 66 Adrian Garrido Sanchis, Barry W. Ninham in an aqueous medium - Activated sludge simulation test., in ISO 11733. 2004, ISO. p. 27. 18. OECD, OECD Guideline for the Testing of Chemicals, in Simulation Test - Aerobic Sewage Treatment: 303 A: Activted Sludge Units - 303 B: Biofilms. 2001, Organi- zation for Economic Co-operation Development: Paris. 19. EEC, Concerning Urban Waste Water Treatrment, E.U. Council, Editor. 1991, EEC: O.J. European Com- munities. p. 13. 20. N.F. Gray, Biology of Wastewater Treatment. 21. S. Massa, M. Caruso, F. Trovatelli, M. Tosques, Com- parison of plate count agar and R2A medium for enumeration of heterotrophic bacteria in natural mineral water, World J Microbiol Biotechnol, 1998, 14(5), 727-730. 22. ATCC, Product Information Sheet for ATCC 15597- B1, A.T.C.C. (ATCC), Editor. 2005, ATCC. p. 2. 23. ATCC, Mehtod 1602: Male-specific (F+) and Somat- ic Coliphage in Water by Single Agar Layer (SAL), U.S.E.P.A.O.o. Water., Editor. 2001. p. 30. 24. A. Garrido, R.M. Pashley, B.W. Ninham, Water steri- lisation using different hot gases in a bubble column reactor, J Environ Chem Eng, 2018, 6(2), 2651-2659. 25. USEPA, Method 1602: Male-specific (F+) and Somatic Coliphage in Water by Single Agar Layer (SAL) pro- cedure: U.S. Environmental Protection Agency Report 821-R-01-029, 38 p. 2001. 26. J. Cormier, Janes, M., A double layer plaque assay using spread plate technique for enumeration of bac- teriophage MS2, J Virol Methods, 2014, 196, 86-92. 27. A.M. Kropinski, A. Mazzocco, T.E. Waddell, E. Ling- ohr, R.P. Johnson, Enumeration of bacteriophages by double agar overlay plaque assay, Methods Mol Biol, 2009, 501, 69-76. 28. P. Zhang, S. Huang, N. Zhang, A.T. Kan, M.B. Tom- son, Automated Analytical Method To Determine Solution Alkalinity of Oilfield Brine in the Presence of Weak Organic Acids, Ind Eng Chem Res, 2019, 58(11), 4667-4673. 29. A. Garrido, R.M. Pashley, B.W. Ninham, Low tem- perature MS2 (ATCC15597-B1) virus inactivation using a hot bubble column evaporator (HBCE), Col- loids Surf B, 2016, 151, 1-10. 30. A.G. Sanchis, M. Shahid, R.M. Pashley, Improved virus inactivation using a hot bubble column evapo- rator (HBCE), Colloids Surf B, 2018, 165, 293-302. 31. M.R.J. Clokie A.M. Kropinski, Enumeration of Bacteri- ophages by Double Agar Overlay Plaque Assay, in Bac- teriophages, U.o. Leicester, Editor. 2009: Humana Press. 32. K. Seo, Lee, J. E., Lim, M. Y., Ko, G., Effect of tem- perature, pH, and NaCl on the inactivation kinetics of murine norovirus, J Food Prot, 2012, 75(3), 533- 40. 33. H. Zhong, K. Fujii, Y. Nakano, F. Jin, Effect of CO2 Bubbling into Aqueous Solutions Used for Electro- chemical Reduction of CO2 for Energy Conversion and Storage, J Phys Chem C, 2015, 119(1), 55-61. 34. W. Knoche, Chemical Reactions of CO2 in Water, in Biophysics and Physiology of Carbon Dioxide: Sympo- sium Held at the University of Regensburg (FRG) April 17–20, 1979, C. Bauer, G. Gros, and H. Bartels, Edi- tors. 1980, Springer Berlin Heidelberg: Berlin, Hei- delberg. p. 3-11. 35. O. Erkmen, Effects of Dense Phase Carbon Dioxide on Vegetative Cells, in Dense Phase Carbon Dioxide. 2012, Wiley-Blackwell. p. 67-97. 36. H.M. Lin, Z. Yang, L.F. Chen, Inactivation of Sac- charomyces cerevisiae by supercritical and subcritical carbon dioxide, Biotechnol Progr, 1992, 8(5), 458-461. 37. R. Orij, M.L. Urbanus, F.J. Vizeacoumar, G. Giaever, C. Boone, C. Nislow, S. Brul, G.J. Smits, Genome- wide analysis of intracellular pH reveals quantitative control of cell division rate by pHc in Saccharomyces cerevisiae, Genome Biol, 2012, 13(9), R80. 38. J.H.S. Richard, C. Flagan, Fundamentals of air pollu- tion engineering. Chapter 2 Combustion fundamentals. Prentice-Hall, Inc.: California Institute of Technology. p. 59 - 166. 39. B.P. Reines, B.W. Ninham, Structure and function of the endothelial surface layer: unraveling the nano- architecture of biological surfaces, Q Rev Biophys, 2019, 52, e13. 40. B.W. Ninham, R.M. Pashley, P. Lo Nostro, Surface forces: Changing concepts and complexity with dis- solved gas, bubbles, salt and heat, Curr Opin Colloid Interface Sci, 2017, 27, 25-32. 41. M.E. Karaman, B.W. Ninham, R.M. Pashley, Effects of Dissolved Gas on Emulsions, Emulsion Polym- erization, and Surfactant Aggregation, J Phys Chem, 1996, 100(38), 15503-15507. 42. H.K. Kim, E. Tuite, B. Nordén, B.W. Ninham, Co-ion dependence of DNA nuclease activity suggests hydro- phobic cavitation as a potential source of activation energy, Eur Phys J E, 2001, 4(4), 411-417. 43. B. Feng, R.P. Sosa, A.K.F. Mårtensson, K. Jiang, A. Tong, K.D. Dorfman, M. Takahashi, P. Lincoln, C.J. Bustamante, F. Westerlund, B. Nordén, Hydropho- bic catalysis and a potential biological role of DNA unstacking induced by environment effects, Pro- ceedings of the National Academy of Sciences, 2019, 116(35), 17169-17174. 44. N.F. Bunkin, B.W. Ninham, P.S. Ignatiev, V.A. Kozlov, A.V. Shkirin, A.V. Starosvetskij, Long-living nano- 67Very High Efficiency of E. coli Inactivation by Body Temperature CO2 Bubbles bubbles of dissolved gas in aqueous solutions of salts and erythrocyte suspensions, J Biophotonics, 2011, 4(3), 150-164. 45. N.F. Bunkin, A.V. Kochergin, A.V. Lobeyev, B.W. Ninham, O.I. Vinogradova, Existence of charged sub- microbubble clusters in polar liquids as revealed by correlation between optical cavitation and electrical conductivity, Colloids Surf Physicochem Eng Aspects, 1996, 110(2), 207-212. 46. W. Kunz, P. Lo Nostro, B.W. Ninham, The present state of affairs with Hofmeister effects, Curr Opin Colloid Interface Sci, 2004, 9(1), 1-18. 47. S. Gudkov, G. Lyakhov, V. Pustovoy, I. Shcherbakov, Influence of Mechanical Effects on the Hydrogen Peroxide Concentration in Aqueous Solutions, Phys Wave Phenom, 2019, 27, 141-144. 48. Z. Fang, X. Wang, L. Zhou, L. Zhang, J. Hu, Forma- tion and Stability of Bulk Nanobubbles by Vibration, ACS J Surf Collo, 2020, 36(9), 2264-2270. 49. N.F. Bunkin, A.V. Shkirin, B.W. Ninham, S.N. Chirik- ov, L.L. Chaikov, N.V. Penkov, V.A. Kozlov, S.V. Gud- kov, Shaking-Induced Aggregation and Flotation in Immunoglobulin Dispersions: Differences between Water and Water–Ethanol Mixtures, ACS Omega, 2020, 5(24), 14689-14701. 50. M. Boström, V.S.J. Craig, R. Albion, D.R.M. Wil- liams, B.W. Ninham, Hofmeister Effects in pH Meas- urements: Role of Added Salt and Co-Ions, J Phys Chem B, 2003, 107(13), 2875-2878. 51. A. Salis, M. Cristina Pinna, D. Bilaničová, M. Mon- duzzi, P. Lo Nostro, B.W. Ninham, Specific Anion Effects on Glass Electrode pH Measurements of Buff- er Solutions: Bulk and Surface Phenomena, J Phys Chem B, 2006, 110(6), 2949-2956. 52. H.K. Kim, E. Tuite, B. Nordén, B.W. Ninham, Co-ion dependence of DNA nuclease activity suggests hydro- phobic cavitation as a potential source of activation energy, Eur Phy JE, 2001, 4(4), 411-417. 53. P. Lo Nostro, B.W. Ninham, A. Lo Nostro, G. Pesavento, L. Fratoni, P. Baglioni, Specific ion effects on the growth rates of Staphylococcus aureus and Pseudomonas aeruginosa, Phys Biol, 2005, 2(1), 1-7. 54. M.H. Adams, Surface inactivation of bacterial viruses and of proteins, J Gen Physiol, 1948, 31(5), 417-431.