I ^| 628 I -065c no.l? ILLINOIS UNIVERSITY, DEPT . OF CIVIL ENGINEERING CIVIL ENGINEERING STUDIES r>> • : "'"' £ The person charging this material is re- sponsible for its return on or before the Latest Date stamped below. Theft, mutilation, and underlining of books are reasons for disciplinary action and may result in dismissal from the University. University of Illinois Library ENGINEER! FEB 2 1 1969 G L161 — O-1096 Digitized by the Internet Archive in 2013 http://archive.org/details/studiesoncoagula17haya y (,& d. COilfiENCE 7 CIVIL ENGINEERING STUDIES SANITARY ENGINEERING SERIES NO. 17 UN SnA, .LL.NO.S 6X801 STUDIES ON COAGULATION EMPLOYING AMMONIUM CHLORIDE AND OTHER AEROSOLS PROGRESS REPORT: June 1, 1962 Through August 31, 1963 By ICHIYA HAYAKAWA CALVIN P. POON TAKEICHI NIREI Supported By DIVISION OF OCCUPATIONAL HEALTH U. S. PUBLIC HEALTH SERVICE RESEARCH PROJECT OH-00127 DEPARTMENT OF CIVIL ENGINEERING UNIVERSITY OF ILLINOIS URBANA, ILLINOIS SEPTEMBER, 1963 PROGRESS REPORT ON RESEARCH GRANT OH-00127 June 1, 1962 to August 31, 1963 STUDIES ON COAGULATION EMPLOYING AMMONIUM CHLORIDE AND OTHER AEROSOLS CHIYA HAYAKAWA, DIRECTOR DEPARTMENT OF CIVIL ENGINEERING UNIVERSITY OF ILLINOIS URBANA, ILLINOIS September 10, 1 963 4 a* ^r SUMMARY The principle findings of Part 1 "THE EFFECTS OF HUMIDITY ON THE COAGULATION OF AMMONIUM CHLORIDE AEROSOLS" are as follows: (1) The weight concentration of air-borne particles with stirring was checked, using Nessler's reaction, revealing a decrease at the rate of a second order reaction, thus showing coagulation to be the main factor in the decrease of weight concentration. (2) Water vapor was shown to decrease the stability of an aerosol of a one percent ammonium chloride solution. With a relative humidity of 50% or less, water vapor affected the coagulation only slightly more. When the relative humidity was 55% or more, water vapor markedly decreased the stability of an aerosol of a one percent ammonium chloride solution. (3) The effect of foreign vapor on the coagulation of an aerosol was evident not only in the case of solid particles, but also with aerosols of 1 iqu i d parti cl es . In addition, Part 2 "SHORT STORAGE STUDY ON THE VIABILITY OF AIRBORNE BACTERIA" has been studied. Data collected to date indicate a very high rate of death during the first 0.5 second of storage. This may be the result of a rapid evaporation of the bound water within the bacterial cells. With a longer storage period, the death rate appears to decrease. Different factors affect the viability of ai rborne bacter i a. Temperature, relative humidity, source of aerosol and growth phase of the bacteria are considered important and preliminary investigations with various combina- tions of these factors have been conducted. Table of Contents Part 1. THE EFFECTS OF HUMIDITY ON THE COAGULATION RATE OF AMMONIUM CHLORIDE AEROSOLS Page I . Introduction 1 II. Formation of Aerosols 3 (a) Chemical Reaction (Condensation Method) 3 (b) Air Atomizer (Dispersion Method) 6 III. Experiment 8 (a) Calculation of Concentration During Inflow 8 (b) Nessler's Reaction Method (Weight Concentration) 10 (c) Calculation of Particle Size 16 IV. Results 22 Bi bl iography 11 Part 2. SHORT STORAGE STUDY ON THE VIABILITY OF AIRBORNE BACTERIA I . ! ntroduct i on 1 (a) Airborne Infection 1 (b) Factors Governing the Survival of Airborne Mi cro-organ i sms 2 (c) Organism Investigated 3 (d) Purpose of the Study 3 II. Theoretical Consideration (a) The Effect of Relative Humidity on the Viability of Airborne Bacteria 5 (b) Protein Structure and Binding Water 5 (c) Denature of Protein 7 (d) Theories of Water Evaporation 7 (e) Evaporation of Aerosols 10 (f) The Effect of Sodium Chloride on the Evaporation of Aerosols 12 III. Experimental Equipment and Procedure (a) Storage Chamber 14 (b) Aerosol Generating Unit 1^ (c) Sampl i ng Uni t 16 (d) Bacterial Culture 16 (e) Experimental Procedures 16 (f) Calculation 17 Table of Contents IV. Determination of the Size of Aerosols V. Results (a) Rate of Death (b) The Effect of Flow Rate on the Viability of Ai rborne E. col i VI . Di scuss i on Bi bl i ography Publication, Staff and Foreign Travel Page 19 20 29 30 38 41 Part 1. THE EFFECTS OF HUMIDITY ON THE COAGULATION RATE OF AMMONIUM CHLORIDE AEROSOLS . Hayakawa T. Ni rei !. INTRODUCTION In these studies the influence of humidity on the coagulation process has been investigated. The values obtained experimentally by others for the coagulation of various aerosols were generally in agree- ment with the value calculated from Smol uchowski ' s theory on the assumption (1 2 3) that the collision efficiency between the particles is 100% . Though some workers have reported a stabilizing effect of a particular vapor on some aerosols, others have found either no effect or a decrease of the stability, for the same system. In an earlier paper , the effects of various foreign vapors (non-polar compounds, neutral polar compounds, and acidic polar compounds) on the coagulation rate of ammonium chloride aerosols were studied. This work provided the background for the present studies involving the effects of humidity on the coagulation rate of ammon ; um chloride aerosols. A tech- nique termed the light scatter decey method, which involves a simplified analysis of the change in light intensity of a Tyndall beam as an aerosol settles under turbulent conditions, was used in both series of experiments. These studies on coagulation employing ammonium chloride aerosols clearly showed that non-polar compounds and neutral polar compounds increased the stability of an aerosol of ammonium chloride whereas acidic oolar com- pounds decreased the stability of the same aerosol. Smol uchowski ' s theory has been shown not to be universally applicable. On the basis of these studies it is felt that additional studies of the effects of humidity on the coagulation rate of ammonium chloride aerosols would be of great value from the practical as well as the theoreti- cal standpoint. The suggestion has been made that a thin layer of vapor or even liquid absorbed on aerosol particles might alter their surface. Smi ronov and Solntseva reported the effectiveness of water vapor as well as butyric acid as aggregants for an ammonium chloride aerosol, but Samokvalov (6) and Kozhukhova stated that either water or octvl alcohol in low con- centrations acted as a stabilizer, rather than as an aggregant, for ammonium chloride. However, Dal la Valle, Orr, and Hinkle reported that water vapor produced the greatest aggregation of such aerosols. The purpose of the studies here»n reported were to investigate whether the coagulation of aerosols was a second order reaction or not, and some of the aspects of aggregation with an emphasis on examining the effects of humidity in the aerosol system. I I FORMATION OF AEROSOLS Consideration is given to two methods by which particulate clouds are formed: (a) condensation method, in which clusters of molecules come together to build up particles of colloidal dimension; (b) desperation methods, in which a substance, initially in bulk or in a state of relatively coarse subdivision, is further split up into fine particles, (a) Chemical Reaction (Condensation Method) In the laboratory, the generation of aerosols by chemical inter- action in the gas phase is very convenient. Certain gases or vapors react chemically with one another to form products which have a very low vapor pressure at ordinary temperatures, e.g., NH.C1 fume from HCl and NhL (Figure 1); H-SO. mist from SO- and water vapor. Since the gases are molec- ularly dispersed, the new particles must first be in a molecularly dispersed condition. The newly formed molecules aggregate and condense to form very fine liquid or solid primary particles. The formation of a mist or fume by the chemical interaction of two or more gases, therefore, ; s essentially a condensation process. The general arrangement of the apparatus for the chemical genera- tion of ammonium chloride aerosol is shown in Figure 1. Dry air filtered th-ough glass wool, is bubbled through concentrated ammonium hydroxide (D) and concentrated hydrogen chloride (C) by the use of an air compressor (A). Gaseous ammonia and hydrogen chloride are mixed in the reaction bottle (E) where the ammonium chloride aerosol is produced. The mixture of gas and aerosol then passes through the water bottles (F) , which act as impingers, and enters the smoke box (S) . Figure 1 SCHEMATIC DIAGRAM OF AEROSOL GENERATOR (CHEMICAL REACTION) A : COMPRESSOR B : FILTER C : HCI (HYDROGEN CHLORIDE) D : NH4OH (AMMONIUM HYDROXIDE) E : REACTION BOTTLE F : WATER BOTTLE G : SODA- LIME S : SMOKE BOX AND STIRRER The water bottles are used for the purpose of "obligatory liquid filtration" . A simple experiment will aid in clarifying the point. Suppose that an aerosol is dispersed by an ordinary atomizer from a dye solution (eosin, fuchsin, toluidine blue, etc.), This aerosol is passed through an impinging series of Erlenmeyer bottles containing distilled water, which was the solvent used to prepare the mother solution. Six to eight of these washing bottles are placed in series and the successive coloration phenomena of the water in the different bottles can be easily observed. The first is intensely colored, the second less, the third still less, and so on up to the sixth, seventh, or perhaps the eighth (according to the dispersion efficiency of the atomizer, the air flow, etc.). There always comes a time when the water in a given one no longer colors. Now if at this moment the aerosol, issuing from the bottle whose water is not colored, is collected with an electric or thermal precipitator, particles whose size is extremely small and of great uniformity are regularly found in the air sample. On this principle three impinging water bottles are used to take out extremely large particles of ammonium chloride aerosols and excess ammonia (or hydrogen chloride;. In this case, the volume of air from the air compressor flows at the rate of 1.0 1/min. and the concentration of the ammonium chloride aero- sol, which enters the smoke box (S) in one minute, is about 11 mg/1. This was checked with Nessler's reaction. Reproduction of aerosol size and concentration is very important for the experiment and physical factors were kept as uniform as possible. The temperature of compressed air, hydrogen chloride, ammonium hydroxide, a reaction bottle and water in water bottles were kept constant and uniform, The pressure of the compressed air must likewise remain constant. A new sample of an ammonium chloride aerosol was made for every experiment. (b) Air Atomizer (Dispersion Method) Since 1920, it has been shown by a number of research workers: that the size of the micellae formed from ordinary atomizers varies w" th the salt concentration of the generating solution, being greater the more concentrated (9) the solution, regardless of the nature of the dissolved material . With the atomizer used by Stalport , for example, the mean diameters of the particles (under an optical microscope) were: 1.6m for 10% solution, 1 .06m. for 1% solution, 0.71m- for 0.5% solution and 0.37^ for 0.1% solution. With an air-liquid jet Lauterback and his co-workers , examining the aerosols it produced, found that when the jet was working above the solution surface, the mass median diameters of sodium chloride crystals were 0.9m for 10% solution, O.^u. for 1% and 0.2u for 0.1%. The median count diameters were respectively 0.08u. for 10%, 0.0V for 1% and 0.03u for 0.1% sodium chloride solutions. When the jet was submerged, that is, when there was some water scrubbing of the large particles, the mass median diameters, for the same solutions, were respectively 0.7m-, 0.3m and 0.2m. In these experiments an aerosol of an ammonium chloride solution was made by the following method, (a) A one percent solution (weight) of ammonium chloride was prepared; and (1 2) (b) Aerosols were created by a Wells-type atomizer under 5 p.s.i.g. air pressure and 5 l/min. air flow. The relationship between the concentration of ammonium chloride solution and the size of aerosol was checked by the light scatter decay method described in my earlier paper . The results were shown in Figure 2. Figure 2 RELATIONSHIP BETWEEN PARTICLE SIZES AND CONCENTRATION OF AMMONIUM CHLORIDE SOLUTION USING A WELLS-TYPE ATOMIZER 10 5.0 82 o "Z -C O c o E E < 0.5 c o o c o u 0.1 e — I / / I 7 / 1 i / 1 / 0.5 1.0 1.5 Particle Radius, //• III. EXPERIMENT (a) Calculation of Concentration During Inflow The number of particles per cubic centimeter for the experiment must be less than 10 particles/cm . The following calculation was applied for the experiments. Nomenclature 3 V: Volume of cloud chamber, 125,000 cm 3 K: Concentration of the flowing aerosol, g/cm 3 Q: Volume rate of flow, cm /min. 3 C: Concentration of the aerosol in the cloud chamber at time t, g/cm t : T i me , min. VpU-QC d£ 0, dt V (K - C) 3 Q cm /min dc -Sdt K g/cm 3 (K-C) V - In (K-C) = ■* t + constant when t = 0, C = 0, constant = - In K ln (K - C) ' V l = e V cm 3 C g/cm 3 3£ Q cm /mi n _£ = e _4 - e ft..*\* I - 1 ♦"» t - ¥ (St) V * 2 K V * In these experiments the Wells-type atomizer atomized approximately one cubic centimeter of fluid of a one percent ammonium chloride solution in ten minutes. Therefore, C-$K t g/cm 125,000 If we assume the concentration of particles was a fifty percent solution, because water evaporates from particles. e ' * l2T!oOO x W " 6,250,000 ' K6 x ,0 " 7 ^ The number of particles per cubic centimeter for the experiments of the light scatter decay method may be calculated knowing that one gram of a one percent ammonium chloride solution was dispersed in a 125 liter cloud chember. Therfore, ? " Too x K53 + Too x ,,0 ° " K26 g/cm3 V - | « r 3 - j x 3.14 x (0.85 x 10"V » 2.57 x 10" ,2 cm 3 where P « density of particle V » volume of a particle r ■ radius of a particle » 0.85u (assuming) W - VP- 2.57 x 10' 12 x 1.26 - 3.24 x 10" 12 g n - ZT - I't* '?n- 12 " °- 4 9 x ,q5 particles/cm 3 W 3.24 x 10"'* 10 (b) Nessler's Reaction Method (Weight Concentration) For observation of particles by the Nessler's reaction method, aerosols of ammonium chloride were prepared with a chemical reaction type aerosol generator (Figure 1), and a rather polydi spersed aerosol was produced. When the supply of aerosol was stopped, the weight concentration = C . After t minutes the weight concentration = C . Figure 3 (a) prepared from experimental data of Table 1 gives non-linear curves for the relation- C o " C t ship between — r and time. Figure 3 (b) , on the other hand, shows a o C q - C t straight line relationship between — r— ~ — and time. This means that the V - c t reaction is of the second order, i.e., — r—r — = K_t as shown below. ° t (15) Because the general second order equation is : A + B *- C + D {j* - K 2 (a - x) (b - x) 1 "•Sfcr-a-K*' a - b A special case of the general second order equation arises when the initial concentrations of both reactants are the same, a = b. This concentration can be purposely arranged in any case, but i t wi 1 1 be necessarily true whenever only one reactant is involved In a second order reaction. When a = b, K ? t is indeterminate. It is best to return to the differential equation, which d * i/ / \2 dx „ . becomes T7 ■ K« (a - x) or , 7T^ = K.,at. dt 2 (a - x) 2 Integration yields -— ■ K t + constant. When t = , x = , so that the constant » 1/a and the integrated rate now is aTT^xJ ■ V c - c In this case, let x » C - C , and a = C,a-x=C then -7-7 - K„t O t O t C C, i o t Aerosols, like most colloidal forms of matter, are essentially unstable, and will usually disappear with the passage of time either by evaporation or precipitation. Evaporation will occur if the substance of TABLE ! RELATIONSHIP BETWEEN WEIGHT CONCENTRATION AND TIME 11 c o mg/100 ml t mi n c t mg/100 ml c - c t C - C. t C C - C t c c. t o 0.245 1 0.198 0.047 0.192 0.970 10 0.130 0.115 0.470 3.620 25 0.095 0.150 0.614 6.450 50 0.058 0.185 0.763 13.200 X 0.492 1 0.412 0.080 0.163 0.396 10 0.195 0.297 0.605 3.100 25 0.110 0.382 0.777 7.060 50 0.071 0.421 0.858 12.050 0.556 1 0.458 0.098 0.176 0.385 10 0.212 0.344 0.620 2.920 25 0.1 14 0.442 0.79^ 6.950 50 0.070 0.480 0.864 12.400 11 Figure 3(a) RELATIONSHIP BETWEEN WEIGHT CONCENTRTION AND TIME ( FIRST ORDER REACTION ) Co-Ct Co 1.0 0.9 0.8 0.7 0.6 0.5 0.4 03 0.2 10 20 30 40 50 60 TIME , MIN. 13 Figure 3(b) C oC, RELATIONSHIP BETWEEN WEIGHT CONCEN TRATION AND TIME (SECOND ORDER REACTION) 20.0 18.0 16.0 14.0 12.0 10.0 8.0 6.0 4 2.0 10 20 30 Time , Min. 40 50 60 14 which an aerosol is composed has an appreciable vapor pressure at room tem- perature. The vapor pressure of a small drop is larger than that from a larger mass of substance, but this effect is not important for a drop radius greater than 0.01 u, C°). Precipitation may occur as a result of diffusion or settling. Aerosols of large particles disappear by settling, and those of very small particles disappear by diffusion. Settling ordinarily accounts for the pre- cipitation of aerosols of radius 0.5 M- or greater. In this experiment par- ticle sizes are 0.5 M- or greater, therefore, diffusion is not important for the instability of aerosols. According to Stokes' law a particle falls with a steady velocity which is proportional to the square of the radius. As the particles come into contact with each other they coalesce, form large particles, or flocculate and precipitate more rapidly than the original particles. A very simple way of increasing the coagulation rate of a particulate cloud is to make the air turbulent by mixing it. Eddies and swirls are then formed and the velocity of the particles relative to each other becomes greater. The chance of collision of particles with one another must therefore be increased, and the observed rate of coagulation will in- crease. On the other hand the mixing will increase the attachment of aerosols on the wall and on the stirrer, Nessler's reagent was used to check the effect of gravitational settling upon the decrease of the mass of the aerosol as follows: After 20 minutes stirring, the weight of ammonium chloride on the wall, on the stirrer, and on the bottom of the cloud chamber respectively was found to be in the ratio 3:3: 9^. Therefore, we can assume the main factor in the decrease of weight concentration depends on gravitational settling. The number of particles settling is given as follows: 15 dn _ n v / ,n dt " h ———————————— --- u; where V: Stokes' velocity (cm/min) of fall for a given particle size and dens i ty . H: The effective height of the box There are two kinds of coagulation in the smoke box; (1) in the air flow the coagulation occurs because of the Brownian motion, and (2) the difference of speed and direction between the small particles and the big particles which moved out of the air flow, results in collisions and coagu- lation. When n equals the number of particles per cc in the smoke box, we can derive two equations from the above two reasons. (1) Brownian motion: dn ... /„■> (2) Collision between small particles and big particles: _ da= K " a(l - a) n — -— (3) Where a: rate of coagulation Then the rate of decreasing particles is expressed from the summation of equations (1), (2), and (3). dn n v dt H K'n + K" (1 - a)n ---(k) In equation (k) , —[—and K'n are independent of stirring and K"a (1 - a) n is affected by stirring. The result of the experiment is a second order equation; and equation (4), which is derived theoretically, also shows a second order reaction. The coagulation of aerosols resembles a bimolecular mechanism, and is described by Smoiuchowski s equation for a second order reaction 16 (C) Calculation of Particle Size The settling rate of a monodi spersed aerosol in a closed chamber under turbulent condi t ions (st i rred settling) is an exponential function of ( 1 8) the concentration, which can be defined as -v t/H (6) n or — = e no where n: the concentration at time t. no: the concentration at time zero. v: Stokes' velocity (cm/min) of fall for a given particle size and density (g/cm ). H: the effective height of the chamber. (19) Equation (2) is applied by Dimmick " to find particle size' from the light scatter decay. The logarithm of the concentration plotted against time re- sults in a straight line, and the same slope is obtained whether n is in terms of number, geometric area, or volume (mass). The slope is defined in terms of the half-life (■§■), a procedure preferred because equation (6) can be v Lj_ transformed to the linear equation: 0.692 = u ? . By Stokes 1 law, H 7 2 v = 7.2 x 10 p r , where p is the specific gravity (density) and r is the particle radius. For convenience a nomograph was made, which relates particle size and density to half-life for a chamber height of 50 cm (Table 2 and Fi gure k) . The fallout of pol ydi spersed aerosols is more complex ' . A reasonable estimate of particle size may be obtained, however, in the following manner. TABLE 2 CALCULATION STOKESIAN RADIUS FROM HALF -LIFE DATA (CHAMBER HEIGHT 50 cm) 17 I Figure h 18 NOMOGRAPH FOR USE IN CALCULATING STOKESIAN RADIUS FROM HALF-LIFE DATA (CHAMBER HEIGHT 50 CM ) HALF- LIFE IN MINUTES 500-r- 400:: 300 200- ioo;i 90 JL 80JL 70JL 60:L 50:: 40JL 30:: 20-- 10 9 8 7 6 5 4 2 — I i PARTICLE SIZE IN MICRONS (RADIUS) 01 I 0-2:l 03— 04 jl 05ir 0-6-; : r 0-7 J L 0-8JL 0.9 -\- I0- L 2-0- 3-Of 40: l 50 ~ 60JL 7.0JL 8-0-ii- 9.0-i l io-oL DENSITY, GM/CC -t-IO-O -i?-9.0 -*0 -I-70 -i*-0 J 13.0 -20 _.|-0 -=■09 -1=0-8 HO-7 ^rO-6 4<>5 ^•4 --0-3 --02 >l 19 As the light scatter decay method is described in an earlier paper by the author, construct a light scatter decay curve and replot on semilog (21) paper . Determine a final exponential slope from the straight portion of the curve which occurs during the later lifetime of the aerosol when most of the larger particles have fallen out. Values on this slope, at about ten minute intervals, are then subtracted arithmetically from equivalent values on the overall decay curve, and the result is plotted on the same scale to yield a second curve line representative of the aerosol less the smallest particles. This second curve is treated as the first; a final slope is determined and subtracted from the second to yield a third, etc. The result is a series of exponen- tial slopes, each representative of an arbitray fraction of the total decay curve, and each having a point at zero time representative of the amount of the light scatter area contributed by that fraction. The particle size of each fraction can be determined from the half-life. The light scatter area contributed by each fraction at zero time is (approximately) a function of the square of the radius of the particles, whereas the mass is a function of the cube of the radius. Multiplying the percent of the light scatter area of each fraction, by its respective size yields the relative mass value of that fraction. These weighted mass values are then totaled, the mass percent calculated, and the cumulative mass per- cent plotted on log probability paper to obtain the median mass radius. For the purpose of this discussion the size distribution of the particles is considered to follow the log probability law and the average radius of the particles, based on either linear radius, cross-sectional area, or volume (mass) are as defined by r., r*, and r,, respectively. The general expression for any average radius whether having a real or hypothetical mean- ing is given by: I 20 m r = (Zn r m /En ) (7) m r r Where m may have the value 1 , 2 , 3 . etc. and n is the number of particles having a radius r. Initially at zero time the total cross-sectional area of particles per cubic centimeter of aerosol is given by, C = n n r 2 (8) o o l Where n is the total initial number. Similarly the initial mass concen- o tration is, ' \ ' V "o ' 4 (9) Where p is the density of the particles. Due to stirred settling the cross-section per cubic centimeter and mass decrease according to equation (6) and may be integrated over the whole particle size range given at time t where u is the standard deviation, C - n r 2 n r e a In 2r (10) •o _ .y__t ^ « iL5 j r 3 n r e H a In 2r (11) Where v is the terminal settling velocity of a particle of radius r. If use is made of the simple Stokes' law, v is found to be equal 7 2 to 7.2 x 10 p r cm/mln, where r is expressed in cm. Taking logarithms of C to base 10, differentiating with respect to t and substituting the value v, equation (10) yields, k H - d( In Ct) = 3 13 x )0 7 P. >o r n r e a In r (12) dt ' H ' ,« _ v-i. f 2 H i ) r n e a In r "o r 21 when t is small compared with H/v, equation (12) becomes r n cr I n r _dil^C*L.3. 13xl0 7 ft _J> [ (13) at n n» 2 r n Jo or In r '0 or _ d(ln C t ) _ . ,, , 7 fi ^ — (1*0 dt 3,,jXlU 'H' 2 r 2 Similarly it may be shown that r 5 - d(l " ^ = 3.13 x 10 7 . g . -| - (15) dt n i r 3 (22) Hatch and Choate integrated the log probability distribution expression to derive the general equation In r P = p In r + 2.303 ^ • I n 2 a (16) m r g i. g Where r is the geometric mean radius and u is the geometric standard deviation. This equation can be applied to transform equat i on s( 1*+) and (15). " !LL 7T* L = 3.13 x 10 7 . ^ . r\ (17) and _ great aggregation is found experi- mentally. This effect is probably due to the fact that substances which lower the vapor pressure of the aerosol constituent can significantly increase the collision probability by the removal of the vapor cushion surrounding a particle. That is, ammonium chloride, being hygroscopic, causes a lower number of water molecules near each particle of an ammonium chloride solution than in the main body of the air, which results in a somewhat reduced pressure in the vicinity of the particles and makes it somewhat easier for two particles to collide. Figure 5 shows the relationship between the relative humidity and the size of an aerosol of ammonium chloride solution. An expression for the equilibrium vapor pressure at the surface of a droplet of solution was derived by Mason . Consider a solution droplet, 23 Figure £ RELATIONSHIP BETWEEN RELATIVE HUMIDITY AND PARTICLE SIZE OF AMMONIUM CHLORIDE 1.2 1.0 0.9 0.8 CO 07 Q < 0.6 LJ —I 05 < fe 0.4 2 0.3 02 0.1 LIQU n U ) s . /CRYSTA L 1 / 1 TEMPERATURE : 21 ± l°C 10 20 30 40 50 60 70 80 90 100 RELATIVE HUMIDITY, % 2U radius r, at the surface of which the vapor pressure P 1 , to be in equi- r librium with an atmosphere in contact with a plane water surface whose equilibrium vapor pressure is Poo. If an element mass dm of water is trans- ferred from the droplet to the plane water surface, the resulting decrease in the free energy of the solution will be: A <(> » 7 ' d A - p dv = ^rf 1 - -^ (19) where dA = (8*Jfr*dr) and dV are the changes in the surface area and the volume of the droplet, 7', p! , and P are respectively the surface tension, density, and osmotic pressure of the solution. Alternatively, the decrease of free energy is given by the work gained in evaporating the mass dm of water at pressure P' , expanding the vapor to the lower pressure Poo and condensing it at pressure Poo, the whole process being carried out reversely and : sothermal 1 y. Assuming the water vapor to behave as an ideal gas, then A4>-*»V '"p^ <») where R Is the universal gas constant, T the temperature, and M the molecular weight of water. Equating the two expressions for &&> gives ln -L = IX* JLM - (21) P w P[RTr p' L R T K ' For solutions whose density vary linearly with concentration (this is very nearly true for NaCl , MgCl,, NH.C1 , etc.), P = R T p l . I n ( 1 + i — ) where n^ is the number of moles of solute dissolved in n. moles of water, P L is the density of water, and i is Van't Hoff's factors which depend upon the chemical nature and the degree of dissociation (i.e., on the concentration) of the solute. When the solution droplet is in equilibrium with the surrounding atmosphere, with its surface temperature equal to that of the air P' must 2$ equal the partial pressure of the water vapor, so H/100 = P|/l^ where H is the relative humidity of the air. Thus the condition for equilibrium becomes: ^L_ 2 y'M P M _ 2 y'M R T P L ,„/.., \ "_ ,n Poo ' p[_ R T r p|_ R T ~ p|_ R T r M ,,n V n 2 ?L R T P r H 2 y'M P L . ,, . n K ,,,\ ln Poo" = ln TOO = p|_ R T r " V L ' ,n + ' ^ " (22) For example: H 2 y'M P L n . n K 100 p' R T r p' n ' 2 7 y' = 73 erg/cm , m = 18 g/mole, R = 8.3 x 10 erg/deg. mole, T = 293°K, Pl = TOO" x ' -53 + 75§" x 1 .00 = 1 . 26 g/cm 3 , P| _ = 1.00 g/cm 3 (assuming the concentration of particles is a 50% solution) r- I x lO-^cm, i =2.7, ^ - ^^ - 0.3 Jj 2 x 73 x 18 1 x In (1 + 2.7 x 0.3) 100 1.26 x 8.314 x 10 7 x 293 x 1 x 10 ' k ' * 26 , -'n 1,81 _-p_,ii . " 1.26 " 1.26 " U ' 4/ 7^5" = 0.625 H = 62.5% Results obtained experimentally by others for the same system show the following: Smi rnov and Solntseva ^ reported the effectiveness of water vapor as well as butyric acid as aggregants for an ammonium chloride aerosol, but Samokhvalov and Kozhukhova v state that low concentration of either water or octyl alcohol act as a stabilizer, rather than as an aggregant for ammonium chloride aerosols. Dal la Valle, Orr and Hinkle ^ , however, report that water vapor produces the greatest aggregation in such aerosols. Many experiments have been carried out on the influence of foreign vapors on coagulation of aerosols, but the literature reveals many cases of apparent disagreement in this field. Some workers have reported a stabilizing 26 effect of a particular vapor on an aerosol; others found no effect or even decreased stability, for the same system. One of the main reasons for this disagreement is that in most of the experiments the sample of aerosol was taken out of the cloud chamber (or smoke box) through a tube, thereby changing the characteristics of the aerosol, because the large particles dropped out as they passed through the tube. The light scatter decay method possesses the advantage that it may be applied to the study of an aerosol without changing the characteristics of the aerosol, because this technique measures directly the change of light intensity of a Tyndall beam as an aerosol settles under turbulent conditions in a cloud chamber. Another reason for the dis- agreement is the fact that the Coagulation rate was not usually determined directly, by measuring the rate of diminution of particle concentration, but was inferred from the settling of the aggregates. (25) Green and Lane concluded that a change in the coagulation rate in the presence of a foreign vapor is possible only in the case of aerosols of solid particles, and that the reason for a variation is not an increase or decrease in the effectiveness of the collisions between the particles, but is due to a change in shape of the aggregates formed. These experiments herein studied show that a change in the coagu- lation rate in the presence of a foreign vapor is also possible w' th aerosols of liquid particles, and that the vapor cushion surrounding the suspended ammonium chloride particles is one of the important factors in the coagulation of ammonium chloride. 27 B! B! LI OGRAPHY (1) R. Whytlaw-Gray and H. S. Patterson "Smoke" Edward Arnold & Co., 184 (1932) (2) D. Sinclair "Stability of Aerosols and Behavior of Aerosol Particles" Handbook on Aerosols, Atomic Energy Commission, 64 (1950) (3) W. B. Kunkel "Growth of Charged Particles in Clouds" J. Applied Physics, 19, 1053 0948) (4) I . Hayakawa "Studies on Coagulation Employing Ammoniun Chloride Aerosols" J. of the Air Pollution Control Association, 12, 265 (1962) (5) L, A. Smirnov and V. H. Solntseva "Conserving the Influence of Ingredient on the Stability of Aerosols" Colloid Journal (U.S.S.R.), 4 , 401 (1938) (6) K. Samokvalov and 0. K. Kozhukhova "Stabilization of an Aerosuspens i on of NH.C1 and Hg 1 " J. of Physical Chemistry (U.S.S.R.) 8, 420, (1936) (7) J. M. Dal la Valle, C. Orr and B. L. Hinkle "The Aggregation of Aerosols" The Physics of Particle Size Analysis, British J. of Applied Physics, H-203 (1954) — — — — -- (8) L. Dautebande "Studies on Aerosols" The University of Rochester, Rochester, N. Y. , 12 (1958) (9) L. Dautebande "Studies on Aerosols" The University of Rochester. Rochester, N. Y., 4 (1958) (10) J. Stalport "Aerosols Medi camenteux IX. De Paction durelique de dinerses administress sous forme d 1 aerosols" Ar C h. Intern. Pharmacodynamic 72, 282 (1946) (11) K. E. Lauterback, A. D. Hayes and H. A. Coelho "An Improved Aerosol Generator" Archives of Industrial Health, 13, 1 56 (1956) (12) W. F. Wei Is "An Apparatus for the Study of Experimental Air-borne Disease" Science, 91 , 1 72 (1940) 28 (13) H. Dessens "The Use of Spider's Thread In the Study of Condensation Nuclei" Quart. J. Roy. Met. Soc . 75, 23 (19^9) (14) ! . Hayakawa "Studies on Coagulation Employing Ammonium Chloride Aerosols" J. of the Air Pollution Control Association , 12, 266 (1962) (15) W. J. Moore "Physical Chemistry" Prentice-Hall, Inc., Second Edition, 535 (1958) (16) H. L. Green and W. B. Lane "Particulate Clouds" D. Van Nostrand Company, Inc., 86 (1957) (17) M. Smoluchowski "Versuch einer Mathemat i schen Theorie der Koagul at i onski net i c Kolloider Losungen" Z. Physikal, Chemle., 29, 129 (1918) (18) D. Sinclair "Stability of Aerosols and Behavior of Aerosol Particles" Handbook of Aerosols, Atomic Energy Commission, Chapter 5 (1950) (19) R. L. Dimmick, M. T. Hatch and J. Ng "A Particle-Sizing Method for Aerosols and Fine Powders" Archives of Industrial Health, 18, 25 (1958) (20) R. I. Sabage "Notes on Sedimentation Models" Washington, Report 1704, National Bureau of Standards (June, 1952) (21) ! . Hayakawa "Studies on Coagulation Employing Ammonium Chloride Aerosols" Je of the Air Pollution Control Association , 12, 266 (1962) (22) H. L. Green and W. B. Lane "Particulate Clouds" D. Van Nostrand Company, Inc., 169 (1957) (23) H. L. Green and W. B. Lane "Particulate Clouds" D. Van Nostrand Company, Inc., 1 70 (1957) (2k) B. J. Mason "The Physics of Clouds" The Clarendon Press, 26, (1957) (25) H. L. Green and W. B. Lane "Particulate Clouds" D. Van Nostrand Company, Inc., 139 (1957) Part 2. SHORT STORAGE STUDY ON THE VIABILITY OF AIRBORNE BACTERIA I , Hayakawa C. P. Poon I. INTRODUCTION (a) Ai rborne Infection In recent decades epidemiological investigations of respiratory diseases such as pulmonary tuberculosis, influenza and the common cold, have established beyond doubt the dissemination of disease by airborn micro-organisms and methods of air hygiene such as disinfection by ultraviolet light and by chem- ical agents have been applied increasingly to their control. One of the most important methods of spread of certain microbial diseases of man is by the expulsion of germ-laden droplets of fluid matter from the human respiratory tract. In sneezing, coughing, and even in talking, great numbers of droplets mostly between 1 and 100 u, in diameter, of saliva and other secretions, some containing micro-organisms, are expelled into the air with considerable velocities . The larger droplets are deposited on nearby ob- jectives or fall to the ground before they can dry, but subsequently they evapo- rate and their residue may be lifted again into the air, as dust, by air currents and mechanical action. Smaller droplets evaporate before reaching the ground, thus leaving any content of micro-organism suspended in the air, called "droplet- nuclei". Such "droplet-nuclei" mav remain airborne for long periods of time and be carried long distances. It is therefore evident that the mechanical transmission of airborne infection depends entirely on the ability of the bacteria to survive while staying in the air. In order to understand the problem of airborne infection. a study of the viability of airborne bacteria seems promising. The inconsistent results of the research work to date also add to the importance and urgent need for further study. ( b) Factors Governing the Survival of Airborne Micro-organisms Factors governing the viability of airborne bacteria are numerous. Among them are temperature; relative humidity (R.H.); particle size, the pres- ence and absence of toxic material, salts and growth medium. Studies in this area have been conducted over the past two decades. The effects of R.H. and temperature on the airborne cells has formed the greater part of these studies, but conclusions drawn from these studies differ widely. Some researchers have stated that the rate of death of the airborne cell is greatest at high R.H. (2 3) levels while others have stated that the greatest rate of death occurs at intermediate R.H. levels . Some reported, on the other hand, that the ( 7 8} lowest R.H. levels are most lethal . More recently, a modern technique of using radioactive cells in aerobi ologi cal studies ' provides a means to distinguish between real and apparent death due respectively to loss of viability of the airborne cells and particulate settling. Some experiments were conducted with different rates of flow in the storage systems and conse- quently apparent environments we-e different. The situation is further com- plicated when airborne cells of different growth phases, in the presence of growth med ; um or without the growth med'um, in the droplets atomized are em- ployed for study. it is obvous that any attempt to compare the data among the workers would be valueless if all existing factors are not taken into cons i derat i on . In order to eliminate the complicated s ; tuat f on, experiments were conducted in different series. The first one. using distilled water in bac- terial suspensions prepared for generating aerosols, served as a control s : nce in this case the only controlling factors were temperature and R.H. Subsequent 3 series were conducted, by adding one or more controlling factors, with or with out eliminating some of the investigated factors. Experiments arranged in this way would show the effect of each of the factors on the airborne bacteria and the role it played in the mechanisms of the death of airborne bacteria. (c) Organism Investigated The choice of suitable bacteria for the study of airborne bacteria is largely dictated by the problem to be investigated. For practical purposes, pathogenic organisms involved in airborne infection are most appropriate for use. However due to the facilities not available in the laboratory for handling pathogens, therefore, it was determined to use non-pathogens in order to eliminate any health hazard possibly occurring to the laboratory personnel. E. col i has the ability to grow in a defined medium. Its growth is reasonably rapid, and it is easy to handle. In addition, it has been studied in other laboratories. Such studies provide a wealth of information and permit a comparison of results. Another virtue of E. col i which makes it more convenient for study is that its membrane is sufficiently permeable . The metabol i cal ly act ; ve centers of the cells are 'n intrimate contact with the environment. This read* ly suggested the ; dea of rapid evaporation of water content of the cells as the cause of death which was the main part of the study in the author's work. ( d ) Purpose of the Study It is the purpose of this study to use E.__col i to investigate the effects of different factors on the viability of airborne bacteria and the mechanism of the death. Radioactive phosphorous P was used to label the bacteria in order to differentiate the physical loss in the storage chambers and the actual death of the organisms. This part of the study, indludlng the short storage study in which a short storage chamber was used to provide storage time ranging from one-half second to four and a half seconds, was primarily for investigating the immediate effects of temperature, R.H., and characteristics of the bacterial suspension on the death of airborne bacteria which was proved, from the present study, to be a result of rapid water evaporation from the cellular material within the bacterial cells. Other factors which do not have immediate effects were not studi ed. it is not the purpose of this study, however, to investigate different kinds of bacteria and to compare the results, but rather to concentrate the author's efforts in a limited time to study on one single species of bacteria. The outcome of which, therefore, can not be applied to all different kinds of bacteria, but it is felt that the result gives a general significance of the effects on airborne bacteria under different environments. II. THEORETICAL CONSIDERATION (a) The Effect of Relative Humidity on the Viability of Airborne Bacteria When a bacteria culture is suspended in distilled water as a spray solution, the osmotic pressure of distilled water is not an important factor in the viability of bacteria because bacteria are remarkably resistant to changes in osmotic pressure and are not disturbed or plasmolyzed by suspension in hypotonic or hypertonic solutions. By taking the osmotic effect out of consideration, the only factors causing damage to the airborne bacteria cells sprayed from distilled water and stored for merely a short time are tem- perature and R.H. since no extrinsic effect by foreign matter is present. Fur- thermore when temperature in the storage chamber is kept within a range in which the bacteria will maintain life, the heat killing in the mechanism of viable decay is ruled out . When R.H. is taken into consideration, it is readily suggested that humidity is closely connected to hydration and dehydration processes, or an exchange of water molecules in and out of the cell membrane. Thus R.H. and temperature, which in turn affects the R.H., have a combined effect on the death of bacteria by removing water out of or in the bacterial cells and the rate of death is governed by the changes of temperature and R.H. in the environ- ment to which the bacteria) cells are exposed. The faster the rate of evaporation of water from aerosol particles is, the faster the rate of bacterial death is ant i ci pated. (b) Protein Structure and Bind'nq Water The primary structure of the protein molecule is the peptide chain (11, 12, 13) built up by L-amino acids. The works of Mellon et a). have sug- gested that the peptide groups can participate in water binding. Perhaps the most evident proof of the presence of water bonded to protein molecules is MM that reported by Fraser and Macrae '. According to their work, investigations of the infra-red spectrum of collagen in the 2 u, region showed that hydrated col- lagen contains a proportion of water molecules which are preferentially oriented with respect to the polypeptide chains. The preferred orientation was suggested by Fraser and Macrae as fol- 1 ows : Tendon Axi s ov A. £ C=0- -H H en o The configuration of the polypeptide chain in collagen is believed to be such that the CO linkages are approximately normal in direction to the collagen chain axis protruding out from the triple chain molecule and so are not available for i nt ra-molecu 1 ar hydrogen bonding as mentioned by Crick and (15) Kendrew . It is possible therefore that the oriented water molecules are linked to these groups through hydrogen bonds as shown above. Identical to th's conclusion are the comments by Li nderstroem-Lang and Sche'lman ' , on the reversible heat denaturation of chymotrypsi nogen . They suggested that water is trapped in the interior of the native protein, formi ng bonds 1 i ke H m ______ o-h ______ oc which on denaturation are transformed into + IhL® 2 There is evidence therefore that water is bonded in native protein molecules as an integral part of the protein structure. Also it is very pos- sible that water molecules are attached to the protein at various sites. The bonds at various sites have different bond strength, some of them being loose bonds and some of them very strong bonds which would require a great amount of energy in order to break them. Bonding sites of water molecules could also determine its relationship to cellular death in such a way that water molecules could be attached to sites not vital to cellular stability and could also be bonded to sites vital to cellular stability. (C) Denature of Protein Among others, one phenomenon in the denature of protein molecules is to remove bonded water molecules from the protein. Since water molecules have been shown to be an integral part of protein molecules, "dehydration" of the protein results in its inactivation which would finally cause the death of the bacter i a. In fact the idea of removal of water molecules within cells causing rapid death is not new. Scott stated that the removal of the most firmly held water molecules results in some loss of bacterial stability. Ferry, ( 1 8} Brown and Damon also suggested that the rate of transmission of water through the cell boundary was involved ; n the death of bacterial aerosols, (d) Theories of Water Evaporation The rate of evaporation of a water drop at rest is governed by the rates of transfer of water vapor and heat between its surface and the environ- ment. The basis of the theory of evaporation of droplets in a gaseous medium was la'd by Maxwell. In the case of stationary evaporation of a spherical droplets Maxwell expressed the rate of diffusion of the vapour of the droplet across any spherical surface with radius r by the following equation: I = $- 4*r D (C - C) (1) dt v oo ' 8 Where Coo ; s the concentration of vapour at infinite distance from the drop with radius r, and C its density or concentration at the surface of the drop which is assumed to be equal to the concentration of saturated vapor at the temperature of the droplet. D is the diffusion coefficient of the vapor and m the mass of a diffusing droplet. !f we assume that the vapour obeys the ideal gas laws and if we express its concentration as the partial vapour pressure p, then M C =RT where R is the universal gas constant and T the absolute temperature. Substituted mto Maxwell's equation, we have _ dm kxr D M (Po - Poo) " dt ' R T ( 2 ) For very small droplets, the evaporation is proportional to changes of surface area, the previous equation can be transformed into: dA 8ji DM ,„ _ v ~^ = -R— (P o " POO) (3) with d the density of the liquid drop. St r ictly speak ng, the evaporation of a droplet can not be a stationary process since the radius and hence the rate of evaporation is constantly decreasing (19) But as shown bv Fuchs when C ^<". d, the evaporation can be regarded as quasi - stationary : .e. the radius can be regarded as a constant value. Fuchs had another approach for developing an equation for water evaporation as reviewed bv Green and Lane . He considered the diffusion process as starting not directlv at the surface of the evaporating sphere but at a distance of A apart from the droplet surface. In other words, the evaporation starts from the surface of an enveloping sphere of radius r +A , where A i s of t h e order of the mean free path of the diffusing molecules. Very few molecules will then be present in the spherical shell of thickness A which represents the distance traveled by an evaporating molecule before it collides with a gas molecule. A can be calculated from the formula m + m 2 1 where \ is the mean free path of the evaporating molecules and m is the mass of an air molecule, m the mass of a diffusing molecule. Considering the equilibrium condition where the diffusing molecules arriving at the surface, r +£ away from the center of the droplet, at a rate equal to the rate at which molecules leave the surface by diffusion, Fuchs finally developed the following equation: ~~ d7 = RTd ^ Po " Pq °) * r y a for very smal ' r ' ^) where a= evaporation coefficient and y = (K T/2m ) 2 , K being the gas constant per molecule. However, when water is evaporated, the heat is removed from the drop and the surface temperature on the water droplet is lower. An equilibrium temperature 8 wi 1 1 be reached later which is lower than the atmospheric tem- perature in the system in which the water droplets are exposed. This depres- sion of droplet temperature (T-6) is significant for water droplet evaporation because a few degrees change would cause a great change in the vapor pressure of the droplet surface. The equilibrium temperature 8 for an evaporating droplet was given by Johnson as: 8 = i-O (f' PT - Pe) (6) K ft T (o/r V a +0 in which the 9 is the equilibrium temperature in absolute degrees, 10 L the latent heat of vaporization of liquid, for water it is equal to 579.5 cal./gram. , K the thermal conductivity of the air,, 0.014 Btu/(cm . sec,) (l°F/in.) or 0.0000288 g. cal/ (cm 2 , sec.) (l°C/cm) , f the relative humidity expressed as a fraction, P T the vapor pressure of the liquid at ambient absolute temperature. p e the vapor pressure of the liquid at equilibrium temperature, D the diffusion coefficient, 0.21 + 0.0015 T°C. ^ 22 > Equation (5) is adjusted with this temperature drop as in the following: dA 8 re M , ,. . ,_ s dF = TTd r V OL. (f.p T - Pf) ) (7) This is the basic equation to be considered in the present study. The previous discussions consider only the case of evaporation of a droplet at a rest or of a droplet wh ; ch has no relative velocity w ; th the movng gas stream. When a droplet is in motion relative to the surrounding gaseous medium ; whether the droplet 's fixed with t h e gaseous stream passing by or the drop'et is moving f'ong w t h t h e gaseous stream at different ve- locities due to the turbulence of t^e flow., a wind factor f should be applied. (23) ^rossllng showed t h 5t t h e w ~d factor \s -. function of Reynold's number for the flow: f = (1 +0.229 R e 5 ; Kinzer and Gunn showed through experiments that f = (1 + 0.22 F Re*) w h ere F should be a function of Re and not a constant as suggested by Frossling. (e) Evaporation of Aerosols Bacterial aerosols are made of various substance differing in physical, chen~:ca! and biological characteristics. They are more complicated than droplets of pure substances in droplet evaporation. In water droplet evaporation, the rate is governed only by the rates of transfer of water vapor and heat between 11 its surface and the environment. !n the case of a bacterial cell, the situation is complicated by two factors. The first one lies in the fact that water mole- cules within the cell are bonded to protein molecules with hydrogen bonds at various sites on the protein molecules. Although there are weak bonds as well as storng bonds of different strength, the average energy required to break down such bonds is much higher than that required to separate the water mol- ecules from free water. The rate of breakage of such protein-water bonds is therefore a limiting factor regarding the water evaporation. The second factor is the presence of the cell membrane whic h attenuates the rate of vapor and heat transfer from within the cell to the outside environment. The cell mem- brane in fact, increases the thickness of the imaginary boundary shell through which a diffusing water molecule *-es to travel before it collides with a gas molecule in the gaseous medium for the completion of the evaporation of this particular water molecule. These two factors., protein-water bonds and cell membrane, are different with var ; ous bacter>a in their effect on the evapora- tion of cellular water. However, the mechanism ; n water evaporation is es- sentially the same in water droplet or in cellular water except that the rate is much slower in the latter case. The equet on express' ng the r ate of water evaporation should apply equally well to both cases, except that in the case of cellular water evaporation, a constant is added to t h e equation to account for the difference of vapor and heat transfer. Therefore we have: ¥ = Constant. |^ r y a (f.P T - Fj (8) dt RT d ' T e For pure water droplets, the constant is equal to unity. The a here is a constant value for the cellular material of one particular kind of bacteria on the average, but no longer equal to 0.04 which is for the case of pure water . It has been shown by Eisner, Quince and Slack that a r e duct ; on of UNIVI 12 this evaporation coefficient a by dispersing a" insoluble agent in water to form insolbule monolayers on sprayed aerosols cuts down the water evaporation by a factor ud to several hundred. It is easily visualized then, that the cell mem- brane would reduce a to a much smeller value 3nd consequently the water evapora- tion from within the cell is greatly reduced. As has been described, the inactivation of bacteria starts when its protein-water bonds are broken and the "freed" water starts to evaporate. The time period for the surrounding water to evaporate to such an extent that the cellular water starts to evaporate is therefore, very important in the study of the viability of ai rbo-ne cells in a very short storage time. From calculation, the water layer surrounding the bacteria evaporates toward com- pletion within a short time. This suggests that the bonded water molecules would be broken almost immediately after the aerosols are sprayed which agrees with t h e immediate lethal effect on the bacteria, (f) The Effect of Sodium Chlor'de on the Evaporation of Aerosols Sodium chloride has the distinct characteristic of lowering the vapour pressure of the liquid medium in w~ c r its ex'sts and consequently reduces the liquid evaporation. Howeve r , the reduced vapor pressure is o" 1 y a few milli- meters of n-ercury for low NaCl concentrat ' ons sucr as the biological saline water used in t h e present studv. The "dehydration ' effect exerted on the bac- terial cells by the surrounding NaCl solution of an aerosol is more important. Since water molecules are leaving the solution at an enormous rate toward com- pletion of evaporation, the concentration of NaCl in the solution is increasing rapidly. As a result, the dehydration takes place because of the higher osmotic pressure. Another possibility of dehydration is that at high salt concentrations, the number of charged groups contributed by the salts is enormous compared to those of protein molecules and therefore more polarizable wate" molecules in the protein a<-e attracted around the salt ions. 13 Furthermore the dehydration of the protein water will start before the complete evaporation of water surrounding the cells. This would not hap- pen in the case of bacterial aerosols sprayed from distilled water, because the NaCl concentration is high enough to attract the cellular water molecules before all water is evaporated. 14 III. EXPERIMENTAL EQUIPMENT AND PROCEDURES (a) Storage Chamber The complete set up of the short storage chamber apparatus is shown in Figure 1 on the following page. The main chamber consisted of a 2 inch I.D. Pyrex glass tube. Nine sampling outlets along the upper part of the chamber were each 7 inches apart so that when the rate of flow in the cham- ber was controlled at 43 1/min., the storage time from one sampling outlet to the next one was half a second. Two dry bulb thermometers placed at both ends of the chamber and a wet bulb thermometer at the outlet end of the chamber served the purpose of R.H. indication. Heat'ng tapes with silicone rubber (Sargent S-40853-2) were wrapped around the chamber for tem- perature control with the aid of autot ransformers (Sargent S-30941 ) . The flow leaving the chamber went through a flowmeter and a recirculation pump, then part of it was exhausted through an exhaust line controlled by a valve. The remaining portion proceeded along the recirculation line through a needle valve 3 for flow control in the system, a bactenal filter and a dehumi di f i er . It then entered the main chamber again, through a three-way opening joint through which humidity controlled air flow and aerosols also entered. One sect'on of the tube between the three-way joint and the main chamber was in- terchangeable. Another piece of lucite tube of 0.875 'nch inside diameter, 5.35 inches long, with five sampling outlets, was used for lower flows other than 43 1/min. These five outlets were 1.27, 1.75, 2.27, 3.08 and 4.30 inches from the inlet end and corresponded to one-half second storage of flows of 10.8, 12.7, 14.4, 17.3 and 21.6 1/min. respectively. (b) Aerosol Generating Unit A Nebulizer No. 40 atomizer was used for generating bacterial aerosols, along with a Devilbiss Compressor 50'. The rate of flow controlled by a screw Figure 1 15 V) TJ X X q 0) O >» >» Nl a a "i a> » m. o > 3 X a> 3 Q 4— o o > 10 a> t- — c 0. v. CL _ o o o c "cvj n E 2 o LL o a E -3 o o (U v_ 4- 3 (J JZ N i_ O (0 CL (0 o 1^ a. o a> a> >4- iZ a> > o .c K UJ o 4- o Q. D 1- «* >» .c o O) o > ^"* o o o in l_ "O 3 a> o> *■ 3 1 c CO 3 (A E 0) 0> a> o E c w a) o 3 O Ix a> c c o > Q 0) l_ 3 a> •*- o o •*- 3 o a> o o a> O 5 CL O CD 2 o or Lu I a. ^COoQLUU-CD 1 _J 5 16 adjustment on the compressor unit was kept at 4.5 1/min, and 2.5 psig. The rate of atomization under this condition was found to be 0.101 ml/min. which was used throughout all experiments in the study. (c) Sampl i nq Un i t Graduated, all glass midget impingers (MSA midget impinger) were used with 10 ml of 0.85% saline water as collecting fluid. A Gast 0211-V36 vacuum pump was used in drawing sampling air from the storage chamber at a rate of 2.6 1/min. controlled by a needle valve in the Matheson flowmeter (Matheson Model 603) . (d) Bacteri al Cu 1 ture E. col i were grown in a defined medium for all experiments: NH.C1 2.0 g Na 2 HP0 £f o.6 g Kr^PO^ — - 0.3 g NaCl --___------ 5.0 g Mg (as MgCl 2 ) 0.01 g S (as Na 2 S0 z+ ) --------- 0.026 g Glucose __ — — — -— ] ,o g (in 100 ml d^st. H 2 0) Distilled water 900 ml P 0. (as H P0, in weak HC1 solution) ------ 10 mc E. coli cultures were harvested in the stationary growth phase throughout all experiments in the study. (e) Experimental Procedures E. col i cultures were harvested and washed three times by repeated centrifugation and resuspension in either one of the following solutions de- pending on which bacterial suspension was to be studied in the storage system: distilled water; 0.85% saline water; 5% saline water and 0.375% glycerol solution 17 Then 0.1 ml of this suspension was removed and was serially diluted in a saline solution and a viable count was determined by the drop plate method using EMB agar. One ml of the same aliquot was placed in a 2-inch diameter aluminum planchet with concentric rings. The R.A. count on this suspension per milli- meter was obtained by placing this planchet with the dried aliquot in a pro- portional counter (NMC internal proportional counter, Model DS-1A, Indianapolis) and counted for 10 min. The suspension was then sprayed as a bacterial aerosol in the chamber by operating the aerosol generating unit. When constant temperature and R.H. was reached at the desired levels, samples at different storage times were taken from the various outlets. Viable counts and the R.A. for each sample were determined in the same way as for the spraying suspension (zero time sample) . (f) Cal cul ati on Let the number of viable cells/0.1 ml of spray suspension = V , and the R.A. counts per minutes/0.1 ml of spray suspension = (R.A.) Q , the V ratio 7 — j-r represents the number of cells per unit radioactive count. The ( R . A . ) q total cells regardless of their viability in each of the samples could be cal- culated as fol lows : o w^ere T = total cells in 0.1 ml of sample (collecting fluid) with storage t i me t , and (R.A.) = R.A. counts per minute/0.1 ml of collecting fluid. The total cells (T ) are comprised of living and dead cells. The number of viable cells per 0.1 ml of collecting fluid can be determined by the drop 18 plate counts, as V . The difference between T and V was the actual death V of bacteria. =— x 100 = percentage of survival after storage time t, or \ 1 - — x 100 = percentage of death of bacteria after storage time t. 19 IV. DETERMINATION OF THE SIZE OF AEROSOLS The size of aerosol particles sprayed from an atomizer depends on a number of things. Among them are type of atomizer used, salt concentration of the generating solution , and jet location relative to the surface of the generating solution in the atomizer * '. In the present study a q bacterial suspension of a concentration of about 2.2 x 10 bacteria/ml was prepared and aerosols were generated with a Devilbiss Nebulizer No. 40 under 2.5 psig and 4.5 1/min. of air flow. The initial size of aerosol particles leaving the atomizer can not be determined directly because water is readily evaporated from the aerosols which makes it impossible to observe directly or indirectly their original s'zes ; f any lapsed time is allowed before the observation. The only means - 9 I Ui > -J UJ or x X cat r- fc s O u. O UJ a: u. o UJ CD I Q. 3 (A O E £ $. o o o ui ui (Z ID 15 Q. UJ Si? UJ o o Q => X UJ > -j a: i o o ^X ' HOD 3 3NU08UIV dO H1V3Q 30 31VH 2U figure 3 THE CHANGE OF RATE OF DEATH WITH TEMPERATURE AT CONSTANT RELATIVE HUMIDITY (E. COLI SPRAYED FROM DISTILLED WATER SUSPENSION) O o UJ 0> c o -O o o 0) o cc 20 1 o o q 7 c I i . I 5 -\ ,30% I >-50 o /J | 4 •* *"r i > 69% < > <^ >> ^ 7i D% < > ^* ^" 2 ^ - . — ■ ** ^ < ^- ■ < < r ^ - ^ ^ ** ** 293 303 3I3 323 333 Absolute Temperature, T° K 25 in Figure k which shows that the evaporation of water droplets changes logarithmically with temperature as in the change of death rate of air- borne E. col i shown in Figure 3. In order to ascertain the combined effect of R.H. and temperature on water evaporation as compared to the death rate of airborne E. col i , kj 2 values were plotted against a function of water evaporation at varying R.H. and temperature as shown in Figure 5. An arbitrary unit was chosen for the function of water evaporation 4- (f.p - p ) \/ /T in plotting Figure 5. Straight lines were determined as being good fits for both cases. Since small water droplets evaporate instantaneously, it was felt that the function plotted in the figure truly represented water evaporation from bacterial aerosols and the instantaneous evaporation is the factor governing the viability of the a>i rborne bacter i a. The presence of salt in the bacterial ae-osols definitely affected the death rate by increasing the death of bacterial cells under the same con- dition of temperature and R.H. except in environments of high temperature and very low R.H. Under this condition, t^e airborne bacter ; a died more rapidly if they were sprayed from distilled water suspension. In order to determine how a change in water evaporation would in- fluence the rate of airborne bacteria, E.__coJ_i aerosols were sprayed from 5% saline and from 0.375% glycerol suspensions. Comparisons of the effects of tne sources of aerosols are shown in Figure 6. it provides the evidence demoiv strating the dependence of bacterial viability on water evaporation from cells Because of the presence of salt in the bacterial aerosols, the bacteria died at a faster rate due to the dehydration effect of the salt and the death rate increased as the concentration of salt in the sprayed suspension increased. In the presence of some hygroscopic material such as glycerol, which impeded 26 Figure 4 THE CHANGE OF RATE OF WATER EVAPORATION WITH TEMPERATURE 283 293 303 313 ABSOLUTE TEMPERATURE, T°K 323 333 UJ O UJ X I- u. o a. x if) H < _l UJ a: ITJ UJ u x »- u. o z o (/) & o (J Ul X I- UJ o (/) Z> O si £2 O LU °.« UJ J^J UJ < u. o or o m o * X UJ o u. o Ul Si 27 21 i a. c o o a. o > UJ w_ o o o cr f X l !l 0D 3 3ujoqjj\/ jo MiDaa j,o 9|Dy 28 Figure 6 COMPARISON OF THE RELATIONSHIP OF THE RATE OF DEATH OF AIRBORNE E. COLI TO RELATIVE HUMIDITY FROM VARIOUS SPRAYED SUSPENSIONS AT 30° C AND VARYING R.K -KM o o Id Ld Z QL O CD rr 5 < X 5 3 Id O U_ 2 O Ld < O 5.00 % Saline Suspension 0.85% Saline Suspension ^ Distilled H 2 Suspension O 0.375 % Glycerol Suspension <^ •i i ■■ C S ~< > 20 30 40 50 60 ( 100- RELATIVE HUMIDITY), % 70 80 29 the water evaporation, the death rate showed a considerable drop compared to bacter ; al aerosols sprayed from a d"stilled water suspension. (b) The Effects of Flow Rate on the V ; a bil ity of A ; rborne E. col i A wind factor "f" was included ; n the equation of water droplet evaporation as suggested by FrossMng to account for the change of evapora- tion rate influenced by the movement of air surrounding the water droplet: dA 8 a D M , , (]2) The wind factor f = (1 + K y Re) w h of sodium chloride was held from increasing and likew ; se the "dehydration" process. Therefore, there was an equ'librium condition when all these factors acted together and the rate of water evaporation from cells under that specific condition was constant. When any one of the factors changed, a new equilibrium wa s formed accompanied by a 3k new rate of death. This change of k values occurred at a slower rate in the case of acosols sal ne suspension origin than that in the case of aerosols of distilled water origin., as shown in Figure 5. The fact that at lower temperatures and high R.H. E. col i aerosols from saline water origin had a higher rate of death than those from distilled water origin can be explained as a result of the "dehydration" process. This process was much less affected by the factors of temperature and R.H. than by the ot^er factors. Although the protei n-bound-wate- evaporation as affected by low temperatures and high R.H. was slow, especially >n the presence of NaCl crystals outside the cell membrane, the "dehydrat ' on" process at high NaCl con- centrations readily took place and this was lethal to the bacterial cells. That was why the rate of death was higher. As the temperature increased and R.H. decreased, the limiting factor seemed to be the combined effect of temperature and R.H. on water evaporation. The consequence showed that in a high tem- perature and low R.H. region, the death rate was higher from distilled water origin. it might be that the effect of NaCl crystals in increasing the bound- ary layer between cells and their outer environment retarded water evaporation at a rate greater than the rate of increasing wcter evapor at' on by the effect of dehydrat 'on. Glycerol in high concentrst ons s tox'c to bacteria, but at a concentration of 0.375% 3 t causes no harmful effect on bacter : a. It is a lipid, highly hygroscopic in nature. When bacterial aerosols were sprayed from glycerol suspension, water was not as readily evaporated as in aerosols sprayed from distilled water suspension. Two different characteristics were encountered in airborn bacteria sprayed fro^ a glycerol suspension: a. Due to its h v groscopic nature, the aerosols would tend to hoi d more moisture onto the cells, so it would take a longer time to evaporate them a 35 In fact it was found in the present study that E. col i suspended in 0.375% glycerol, as examined by a Carl Zeiss phases microscope, showed an average size of somewhat between 1 and 2 u., while the same culture suspended in dis- tilled water showed an average size of less than 1 u.. This incidated that in glycerol suspension, E. col i had a higher water content. b. The glycerol as a lipid material, would reduce the rate of water evaporation. It is known that in the presence of impurities in water particles, the rate of water evaporation is retarded. The same phenomenon occurred here. If we rearrange Equation (7) and integrate, we obtain My a (f.P T - pj (,3) This is an equation for calculating the time required to evaporate a water droplet with radius r to a size of r . Anything which reduces the a value would result in a longer time for the evaporation of the droplet. The glycerol here probably played the same role in reducing the a value as the (29) fatty alcohol did as reported by Eisner, Quince and Slack . The phenomenon of abrupt change of the rate of death after 0.5 seconds was not very well understood. While a great portion of the population of bacteria died off in the first half second of storage, how the remaining portion maintained life is surprising. It is felt that this phenomeon has to do with the protein-water bond. Protein water can be linked to protein molecules through hydrogen bonds at various sites. Some of them link to the exterior of protein molecules whereas others could be trapped in the interior of the native protein. Some of them are possibly vital and some of them probably not. Some protein molecules might not even have water molecules in them. Chances were a great portion of the bacteria had lost the protein water which was vital and died off during that 0.5 seconds of storage and yet the remaining bacteria had lost only non-vital bonded water or had no bonded 36 water to lose at all. Those that managed to survive had only protein water which was strongly bonded and trapped in the interior of protein molecules. These water molecules were very difficult to evaporate not only because of the strong bond but because of their position inside the protein molecules which provided them with good protection against the process of evaporation. The rate of evaporation was therefore very slow and consequently the rate of death was relatively slow compared to that in the first 0.5 seconds of storage. In fact some of the bacteria could survive a long time as reported by others, probably because some protein-water was held inside indefinitely without evaporation. Based on this explanation, the sudden change of the rate of death during the one half second of storage period suggested that within this period of tir^e, both the free water outside the bacterial cells and the bonded water linked to the exterior of the protein molecules was mostly evaporated. Never- theless the exact time when the bonded water molecules started diffusing out was not known. Since the lethal process actually started sometime within th : s one-half second storage time, the k, values could be larger if they were 2 calculated based on the exact lethal process time period instead of the one- half second storage time. For this reason, the k ± values reported in the 2 present study did not represent the absolute death rate but instead gave the 3Dpare->t death rate of th ; s half-second storage time. It would seem that to compare the k values with the other research workers is impractical since no attempt has been made by any workers to find out the exact time when the rapid death rate did change to the relatively slow rate. In order to find out this exact time, samples at one-tenth of a second intervals or even shorter should be taken. However, the design and precision of instrumentation would make this study very difficult at the present time. 37 The death rates were postulated to follow the first order reaction in short storage. However, it was felt that this was not necessarily true for all cases. Many workers have also postulated partial order reactions from f'rst to fifth order reactions. The fact is that in the study of bio- logical systems using death or certain inactivation as the end point for rate determination, the same reaction could go on in different parts of a protein molecule at different rates. The breakage of one protein-water bond may cause death or protein inactivation whereas the breakage ot one protein-water bond in another part of the same molecule may not be detrimental. Therefore the lethal process could proceed along different paths to reach the same end poi nt. To compare the apparent death rate of the short storage, Webb's (8) data was used. However no half second sample was collected in his study, the whole first second storage period in the present study must be treated as an exponential function so that it was possible to compare results with Webb. At 25°C and 50% R.H., the k for E. col i as calculated from Webb's data was 1.52. In the present study, k at 20°C and 50% R.H. was 1.14 and at 30 C and 50% R.H. was 1.35. The interpolated k. at 25°C and 50% R.H. was 1.24 which was slightly smaller than that which Webb has found. The study of velocity of the aerosol flow in the environment indicated no noticeable effect on the viability of bacterial aerosols within the range of velocities investigated. With wind factor f = (1 + K V"'Re) , a particle of (24) 13 n gives a K value practically zero from Kirlzer's experimental curve and the wind factor becomes practically unity. 38 BIBLIOGRAPHY (1) J. P. Ouguld "The Size and the Duration of Air-carriage of Respiratory Droplets and Droplet-nuclei" Journal of Hygiene Cambridge. 44. 471. (1946) (2) K. B. DeOme "Effect of Temperature and Humidity and Glycerol Vapor on Viability of Airborne Bacteria" Am. Journal of Hygiene . 40, 239-250, (1944) (3) C. G. Loosll, H. M. Lemon, 0* H. Robertson and E. Appel "Experimental Airborne Influenza Infection" Proc. Soc. Exptl. Biol. Med . 53, 205-210, (1943) (4) E. W. Dunklin and T. T. Pick "The Lethal Effect of Relative Humidity on Airborne Bacteria" J. of Exptl. Med .. 87, (1948), pp. 87-101. (5) R. M. Ferry and L. G. Maple "Studies on the Loss of Viability of Stores Bacterial Aerosols" J. of Infectious Diseases . 95, (1954) pp. 142-159. (6) L. L. Shechmeister and L. J. Goldberg "Studies on the Epidemiology of Respiratory Infections" J. of Infectious Diseases . 87, (1950), pp. 116-127. (7) W. Wells and P. Zapposont "The Effect of Humidity on B-streptococci Atomized in Air" Science . 96, (1942) pp. 277-278. (8) S. J. Webb "Factors Affecting the Viability of Airborne Bacteria" Canadian J. Microbio . . 5, (1959), pp. 649-669. (9) G. J. Harper^, A. M. Hood and J. D. Morton "Airborne Microorganisms: A technique for Studying Their Survival" J. Hyg.. 56, (1958), pp. 364-370. (10) R. B. Roberts, D. B. Cowle, P. H. Abelson, E. T. Bolton and R. J. Britten "Studies of Biosynthesis In Escherichia coll" Carnelgie Institute of Washington Publication, 607, Washington, D. C. , (1957), pp. 58-91. (11) E. F. Mellon, A. H. Karn and S. R. Hoover "Water Absorption of Protein: 1. Effect of Free Amino Groups in Casein" J. Am. Chem. Soc . v. 69. (1947) pp. 827-831. (12) E. F. Mellon, A. H. Karn and S. R. Hoover "Water Absorption of Proteins: Effect of Physical Structure" J. Am. Chemical Soc . v. 71, (1949), pp. 2761-64. 39 (13) E. F. Mellon, A. H. Karri, and S. R. Hoover "Water Absorption of Proteins: Effect of Guanidino Groups of Casein" J. Am. Chem. Soc . , v. 73, (1950, pp. 1870-71. (]k) R. D. Fraser and T. D, Macrae "Possible Role of Water in Collagen Structure" Nature, 183, (1959), pp. 179-180. (15) F. H. C. Crick and J. C. Kendrew Adv. in Protein Chem . , v. 12, (1957), p. 133. (16) K. U. Li nderstrom-Lang and J. A. Schellman The Enzyme , Protein Structure and Enzyme Activity , v. 1, Chap. 10, (New York (1959) , p. ^90. (17) W. J. Scott "The Effect of Residual Water on the Survival of Dried Bacteria During Storage" J. Gen. M'crobio ., v. 19, (1958), pp. 624-633. (18) R. M. Ferry, W. F. Brown and E. B. Damon "Studies of the Loss of Viability of Stored Bacterial Aerosols: 11. Death Rate of Several Non-Pathogenic Organisms in Relation to Biological and Structural Characteristics" J. of Hyg ., v. 56, (1958), p. 125. (19) N. A. Fuchs Evaporatio n and Droplet Growth in Gaseous Medi a, (New York, 1959) » Chap. Ill, pp. 60-66. (20) H. L. Green and W. R. Lane Particulate Clouds: Dust, Smokes and Mists , New York, p. 8k. (2P J. C. Johnson "Measurement of the Surface Temperature of Evaporating Water Drops" J. of Applied Physics , (1950) 21 (1), 22. (22) J. Hilsenrath and Y. S. Toulaukian Tran. Am. Soc. M ech. Engr ., 76, 967, (195*0 (23) N. Frossling "Uber die Verdunstung Fal lender Tropfen" Beitr. Geoph ys., v. 52, (1938), p. 1 70 (2k) G. D. Kinzer and R. Gunn "The Evaporation, Temperature and Thermal Relaxation Time of Freely- Falling Water Drops" J. Met ., v. 8, (195O , p. 71. (25) T. Alty "The Maximum Rate of Evaporation of Water" The Philosophical Magazine , (1933) ser. 27, 15, p. 82. ko (26) L. Dantebande "Studies on Aerosols" University of Rochester, k, (1958). (27) J. Stalport "Aerosols Medi camenteus IX. De L'action Durelique de diverces substances administress sous forme d'aerosols" Arch. Intern. Pharmacodynami e , v. 72, (1946) , p. 282. (28) K. E. Lanterbach, A. S. Hayes and M. A. Coelho "An Improved Aerosol Generator" Arch. Ind. Health , v. 13, (1956)., p. I56. (29) H. S. Eisner, B. W. Quince and C. Slack "The Stabilization of Water Mists by Insoluble Monolayers" Discussion Faraday Soc. 30 , 86., (i960) PUBLICATION, STAFF AND FOREIGN TRAVEL A. Publ (cations "Studies on Coagulation Employing Ammonium Chloride Aerosols" J. of Air Pollution Control Association , 12, 266. (1962) "The Effects of Humidity on the Coagulation Rate of Ammonium Chloride Aerosols" (in preparation) "Short Storage Study on the Viability of Airborne Bacteria" (in preparation) B. Staff Ichiya Hayakawa, Principal Investigator - Time as required, June, 1962 - present Calvin P. Poon , Research Assistant - Half time, June, 1962 - June, 1963. Full time ; July, 1963 - to present Date of Birth: November 8, 1935 Country of Birth: Canton, China Present Citizenship: Chinese Sex : Ma 1 e Education Experience: Degree insti tute Conferr i nq Field Year B.S. National Taiwan University Civil Engr. 1958 M.S. University of Missouri Sanitary Engr. I960 Research Background: Insti tution Nature Year University of Illinois Petrochemical Waste Treatment 1960-61 University of Illinois Radon Gas Emanation from Soil 1961-62 Surface 42 inst i tut ion Nature Year University of Illinois Coagulation of Aerosols 1962-present University of Illinois Factors Affecting 1962-present the Viabi 1 i ty of Ai r- borne Bacteria Takeichi Nirei, Research Assistant - Full time, July, 1962 - August, 1962. Half time, September, 1 962 - June, 1963. Full time, July, 1963 - August, 1963. Date of Birth: September 27, 1 930 Country of Birth: Tokyo, Japan Present Citizenship: Japanese Sex: Male Education Experience: Degree Institution Conferring Field Year B.S. University of Chiba Architectural Engr. 1953 Research Background: I nsti tution Nature Year Kanagawa Technical High School Architecture 1953-62 University of Illinois Coagulation of 1962-63 Aerosol s C. Foreign Travel There has been no foreign travel associated with this research grant by August 30, 1963. / UNIVERSITY OF ILLINOIS-URBANA R7RII65C C001 CIVIL ENGINEERING STUDIES. SANITARY ENGI 17 1963 3 0112 008520774