Graphical Procedurefor Comparing Thermal Death of Bacillus stearothermophilus Spores in

Graphical Procedure for Comparing Thermal Death of Bacillusstearothermophilus Spores in Saturated and Superheated Steam

JAMES J. SHULL' AND ROBERT R. ERNST

Wilmot Castle Company, Rochester, New York

Received for publication April 17, 1962

ABSTRACT

SHULL, JAMES J. (Wilmot Castle Co., Rochester, N. Y.)AND ROBERT R. ERNST. Graphical procedure for com-paring thermal death of Bacillus stearothermophilus sporesin saturated and superheated steam. Appl. Microbiol.10:452-457. 1962-The thermal death curve of driedspores of Bacillus stearothermophilus in saturated steamwas characterized by three phases: (i) a sharp initial risein viable count; (ii) a low rate of death which gradually in-creased; and (iii) logarithmic death at maximal rate. Thefirst phase was a reflection of inadequate heat activationof the spore population. The second and third phases rep-resented the characteristic thermal death curve of thespores in saturated steam. A jacketed steam sterilizer,equipped with a system for initial evacuation of thechamber, was examined for superheat during normal op-eration. MNeasurements of spore inactivation and tempera-ture revealed superheat in surface layers of fabrics beingprocessed in steam at 121 C. The high temperature of thefabric surfaces was attributed to absorption of excess heatenergy from superheated steam. The superheated steamwas produced at the beginning of the normal sterilizingcycle by transfer of heat from the steam-heated jacket tosaturated steam entering the vessel.

*While participating in the development of advanceddesigns for steam sterilizers, we found it necessary todevise an accurate method for measuring the effect ofsuperheated steam on the thermal inactivation of bacterialspores. Henry (1959), Knox, Penikett, and Duncan (1960),and Bowie (1961) showed that superheating occurs insteam sterilizers because of the combined effects of heattransfer from steam to the fabric and hydration of dryfabrics by steam. The amount of superheat is inverselyrelated to the moisture content of the fabric to be sterilized.Walter (1948) observed superheat, during sterilization ofsurgical dressing packs, resulting from absorption of heatfrom the hot walls of a jacketed chamber prior to the in-troduction of steam. Savage (1937) found that smallamounts of superheat may be tolerated in a sterilizingsystem, but, as the deviation from the phase boundaryincreases, the rate of spore destruction drastically de-

' Present address: Department of Bacteriology and Botany,Syracuse University, Syracuse, N.Y.

creases. Consequently, the amount of superheat obtainedin sterilizers is an important consideration in their design,and the preferred method for detection of superheat isthrough use of bacterial spores.

It is difficult to reproduce steam environments in thechamber of a steam sterilizer. Therefore, analysis of asingle sterilizing cycle for its sporicidal effectiveness wasmandatory in the present study. We can predict the num-ber of spores surviving an abbreviated steam treatmentfrom time-temperature records of that treatment. Then,with the thermal death curve of the test organism insaturated steam as a reference, we can detect the presenceof superheated steam as it is manifested in a greater sur-vival of spores than predicted.Thermal death times and thermal death time curves

have been established for dried bacterial spores in satu-rated steam (Kelsey, 1958), but the determination of theshape of the thermal death curve of spores in steam hasmet with technological difficulties. The problems involvedrelate mainly to the control of steam. They include: main-tenance of steam temperature and pressure at the phaseboundary; accurate, uniform, and reproducible tempera-ture control; rapid heating and cooling and necessary cor-rections; and the relatively high reaction rate. Rapid heat-ing is difficult unless a hot jacket is maintained about thechamber (Walter, 1948; Fallon, 1961). Our results showthat a hot jacket superheats incoming steam. Contact ofsuperheated steam with the spore samples can be avoidedby placing them within a bundle of fabric. The super-heated steam entering the bundle loses its excess heat tothe surface layers of the fabric, and its temperature dropstoward that of saturated steam. The moisture content ofthe fabric is critical to the establishment of saturatedsteam environments in such systems (Henry, 1959; Knoxet al., 1960). In the present investigation, normally hy-drated cottons (5 to 6 % water by weight) were used. As aresult, superheat of 1 or 2 C may have been present in ex-periments concerning "saturated steam." This smallamount of superheat, if present in our system, may havebeen responsible for scattering of experimental points, butdid not affect the major characteristics of the thermaldeath curve.

If we place spore controls within a bundle of fabrics, weintroduce another complicating factor, as displacement ofair by incoming steam is retarded by fabrics (Knox and

452

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/a

m o

n 19

Nov

embe

r 20

21 b

y 15

3.13

6.10

9.21

2.

THERMAL DEATH OF SPORES IN STEAM

Penikett, 1958; Fallon, 1961). The recent development of"high-prevacuum" sterilizing systems by Knox andPenikett (1958) provided a solution to the problem. Theseworkers reported that evacuation of a chamber to below23 mm of Hg absolute pressure prior to introduction ofsteam permits instantaneous penetration of the steam intothe interstices of tightly packed bundles of fabrics. Usingthis procedure, we obtained rapid heating of sampleswithout subjecting them to superheated steam.The procedure normally used for determination of

thermal death curves involves exposure of replicate sporepreparations to identical conditions for varying timeperiods. These procedures were briefly reviewed byHumphrey and Nickerson (1961) and Stern and Proctor(1954). Because it is difficult to reproduce steam systemsaccurately in a sterilizer chamber, we employed the alter-nate process of comparing separate determinations to acommon reference. This was done by the following modifi-cation of the graphical method of Bigelow (1920) fordetermining steam-process times for canned goods.Assuming that the thermal death curve of a micro-

organism is truly logarithmic, we established death rateconstants in accordance with the expression

kT = -(log No- log N) (1)

where No is the initial population, Nt is the populationat time t, and k is the death rate constant at temperatureT. We can determine thermal death time curves forbacterial spores in steam with fair accuracy. Since, atthe thermal death time (TDT) of the spore population,the remaining viable population numbers less than onespore, log Nt can be taken as zero, and the expressionsimplified to

kT = -log NolTDT (2)Values for k obtained from thermal death time data

plot logarithmically against the reciprocal of absolute

Temperature .6

120

o ~~~~~~~~~~~~.4 E

0110 -kd1..~ ~ ~~~00

E.2

100 0U 1 2

Time - minutesFIG. 1. Graphical prediction of thermal death of bacterial spores

in steam.

temperature over a limited range. By use of this informa-tion, a time-temperature record of steam treatment of thespore preparation can be converted graphically to a time-death rate curve (Fig. 1). If the assumption of logarithmicdeath is correct, the area under this curve will be directlyrelated to the number of logarithms by which the sporepopulation was reduced, or

I -kdt = log No- log Nt0

(3)

Comparison of predicted survival of a spore populationin a series of determinations by the above procedurewith plate counts of surviving spores should readilyreveal the true shape of the thermal death curve. Onecan illustrate this graphically by the procedure describedin Fig. 2. A straight line of any negative slope is drawnfrom a point on the ordinate representing the log of theinitial population of spores. This is the common referencereferred to previously. The Y coordinate for any givensteam treatment is established by subtracting Jfot-kdtfrom the initial population. The corresponding X co-ordinate on the reference curve is the arbitrary time co-ordinate of the experimentally established final sporepopulation.The procedure just described may be used for detection

of superheat during the operation of a sterilizer. In thiscase, the time-temperature record of steam treatment ofspores placed at the site of interest is used to predict theirsurvival, assuming that saturated steam conditions pre-vailed. If superheated steam were present at the site,the actual survival would be greater than predicted fromthe thermal death curve of spores in saturated steam.

MATERIALS AND METHODSThe test organism used was Bacillus stearothermophilus

1518 obtained from Z. J. Ordal, University of Illinois,Urbana. Resistant spores were produced by growth ofthe organism from a heat-shocked inoculum on Nutrient

~~~~ ~~ActualZ AI'*SuCrvivalI

.> -kdt 1I

Time - arbitraryFIG. 2. Graphical representation of predicted and actual survival

of spores in steam.

1962] 453

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/a

m o

n 19

Nov

embe

r 20

21 b

y 15

3.13

6.10

9.21

2.

SHULL AND ERNST

Agar containing 0.0001 % MnSO4. H20. The cultures wereincubated at 55 C for 48 hr. The sporulated culture washarvested and washed three times in distilled water bycentrifugation. The resulting pellet was resuspended indistilled water and the suspension was heated at 100 Cfor 5 min. This stock suspension was diluted to the de-sired level, and 0.01-ml portions were placed on bibulouspaper strips. The strips were dried in desiccated air at45 C. Thermal death times in saturated steam, determinedfor the spore preparation by other members of this lab-oratory, were used to establish the death rate constantsdiscussed above. The thermal death time for a populationof 1.3 X 106 spores in saturated steam at 121 C was 13min.

Spore preparations were exposed to steam in a Castle2036 rectangular jacketed sterilizer (Wilmot Castle Co.,Rochester, N.Y.) equipped with controls manufacturedby Drayton Regulator Co., Ltd. (West Drayton, Middle-sex, England). These controls permitted rapid evacuationof the sterilizer chamber to below 17 mm of Hg absolutepressure, thus providing instantaneous penetration ofsteam into textile and paper materials within the cham-ber. A high rate of steam delivery to the chamber wasprovided with facilities for final evacuation of the cham-ber for rapid cooling and drying of fabrics containedtherein (Penikett, Rowe, and Robson, 1958). Six to tenspore-impregnated strips were employed in each experi-ment. For saturated-steam studies, these were placed inthe center of folded linen towels packed in a 10-in. cubicalpaper carton. In experiments related to the detection ofsuperheat, the contaminated strips were placed at thesuspected site. A 32-gauge iron-constantan thermocouplewas placed with the spore strips and led through thegasket seal on the sterilizer door to a potentiometricrecorder (Bristol Co., Waterbury, Conn.). A referencejunction was placed in saturated steam at atmosphericpressure. The emf generated by the thermocouple wasrecorded on a scale (0 to 20 mv), and the readings ob-tained were converted to temperature.The chamber containing the spore preparation was

evacuated to 17 mm of Hg absolute pressure. Steam wasthen admitted to the chamber until the operating pressurewas attained, and the system was held constant for thedesired time interval. The steam was then released, thechamber was evacuated to cool and dry the materials,and sterile air was admitted to the chamber to completethe cycle. The time-temperature record was converted toa time-death rate curve, and the area under the curvewas determined with a planimeter. This value was usedto determine the expected survival of spores on thestrips.The spore strips were removed from the chamber,

ground in sterile distilled water in a Waring Blendor, andplated by the standard plate-count procedure (AmericanPublic Health Association, Inc., 1953) with the followingmodifications. Distilled water was used as the diluting

fluid and Dextrose Tryptone Agar (Difco) was employedas the plating medium. Petri dishes received a shallowlayer of sterile agar medium prior to plating and wereoverlaid with sterile agar after plating to minimize spread-ing of colonies. The plates were incubated in a humidifiedincubator at 55 C for 48 hr. The spore populations ofuntreated controls were determined by the same procedureexcept that a 10-ml sample of the first dilution in eachcase was heated to 100 C for 5 min, and further dilutionsfor plating purposes were made from this sample. Thisprovided the initial "heat-activated" count of the sporepopulation. When the lower end of the thermal deathcurve was tested, removal of the spores from the stripswas effected by placing the strips in small volumes ofdistilled water and subjecting them to agitation in anultrasonic field. The entire sample was then plated ineach case.

RESULTS

Thermal death curve of B. stearothermophilus spores insaturated steam. Plate counts of 31 control spore-impreg-nated paper strips revealed an arithmetic mean of 1.3 ±t

I1\i4

I\\ Reference Curve

3

z

0-J

0 Time-arbitraryFIG. 3. Thermal death curve of Bacillus stearothermophilus spores

in steam.

454 [VOL. 10

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/a

m o

n 19

Nov

embe

r 20

21 b

y 15

3.13

6.10

9.21

2.

THERMAL DEATH OF SPORES IN STEAM

0.2 X 105 spores per strip. The results of 22 experimentsinvolving steam treatment of the spores are recorded inFig. 3. The vertical line through each point representsthe standard deviation from the arithmetic mean of theplate counts of five to ten individual spore strips. Thecurve was established by a least squares fit of the mean

populations.The response of the "viable" spore population to sat-

urated steam as represented by the thermal death curve

was characterized by three phases: (i) a sharp increase inviable spore count after brief heating in steam; (ii) a

low rate of death, gradually increasing upon continuedheating; and (iii) a logarithmic order of death at maximalrate. The logarithmic portion of the curve was of greaterslope than the reference curve, crossed the latter, andextrapolated to extinction at a time prior to the equivalentthermal death time of the spore population.

Superheat in sterilizers. In an attempt to establish theextent, cause, and effect of superheat in surface layers offabric bundles during routine sterilizer operations, we

made a series of measurements of the temperature charac-teristics of the loaded and unloaded sterilizer chamber.The jacket temperature was maintained at the operatingtemperature (132 C) for the duration of the sterilizingcycle. The time-temperature record of steam at thegeometrical center of the empty sterilizer is reproduced inFig. 4. Extremely high temperatures were observed for a

short period upon entry of the steam into the sterilizer.Within 1 min, the temperature dropped to the intendedoperating temperature and remained constant until theend of the sterilizing cycle.The effect of the superheated steam on the surface

temperature of fabrics was investigated by placing thespore strips with a thermocouple in the surface layers of

linens in the open-topped 10-in. cubical carton. Fig. 5 isthe time-temperature record of such an experiment com-

pared with that for the center of the linens in the same

system. The surface layers were superheated by 6 Cabove the temperature of saturated steam recorded atthe center of the bundle. This condition remained stablefor the duration of the sterilizing cycle. When the time-temperature records of the surface layers were employedto calculate reduction in numbers of spores on the basisof saturated steam, the resulting figures were markedlygreater than the actual reduction measured by platecounts (Table 1). The amount of superheat in theseexperiments was not definitely known owing to the in-accuracy of the available pressure-measuring equipment,and comparison of results obtained from separate ex-

periments is therefore invalid. Measurements of saturatedsteam temperature in separate cycles with the same systemindicated that superheat did not exceed 8 C in any of theexperiments.

127 Surface

121 / t ~~~CenterlII21

E

Time - minutesFIG. 5. Temperature record of fabrics in a "high-vacuum" steam

sterilizer.

2

Time- minutesFIG. 4. Behavior of steam temperature in an empty steam-jacketed

sterilizer.

TABLE 1. Influence of superheat on the thermal death of bacterialspores under practical sterilizing conditions

log No - log Nt*

Experiment Initialnumber Population Assuming saturated steam

T019 No) ~ -Experimentalf t_kdt Predictedt

1 5.11 4.76 >5.11 1.542 5.11 4.45 >5.11 0.253 5.11 5.53 >5.11 1.654 5.11 6.75 >5.11 0.235 5.11 8.6 >5.11 2.11

* No = initial population; Nt = final population after exposureto superheated steam.

t Values determined graphically by relating fo0-kdt to thethermal death curve in Fig. 3.

1962] 455

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/a

m o

n 19

Nov

embe

r 20

21 b

y 15

3.13

6.10

9.21

2.

SHULL AND ERNST

DISCUSSION

Thermal death curve. Several investigators have demon-strated an initial delay in the thermal death of spores ofthermophiles (Stern and Proctor, 1954; Humphrey andNickerson, 1961). Halvorson (1958) suggested that suchcurves were probably the result of interaction of twophenomena. The first of these is heat activation of thespores, resulting in increased percentage germination. Thesecond is the hypothetical multiple-hit phenomenon, ac-cording to which several critical spots or areas within eachspore are inactivated in the death process. The methodsemployed in this study appear to have partially separatedthe phenomenon of heat activation from the true thermaldeath curve. This resulted first in an increased viablecount, reflecting the influence of heat in stimulating ger-mination. Heat activation required a relatively short timeat the temperatures employed, and was apparently com-plete long before the end of the slow death phase of thethermal death curve.Attempted heat activation at 100 C for 5 min was

insufficient to provide 100 % germination of the sporepopulation. The control population so treated gave riseto mean plate counts of 1.3 + 0.2 X 105 spores per paperstrip. Extrapolation of the initial portion of the thermaldeath curve to zero time resulted in a hypothetical initialpopulation of 2.4 X 105 spores per strip, an increase of80 % over the controls. It is likely that the low initialdeath rate observed in the thermal death of B. stearo-thermophilus is real rather than an artifact resulting frominsufficient heat activation of the initial spore population.Under certain conditions of experimentation, however,sharp delineation between heat activation and true thermaldeath curve may not occur (Humphrey and Nickerson,1961), resulting in confusion of one with the other.

Superheat in steam sterilizers. Examination of the char-acteristics of operation of an empty jacketed steamsterilizer revealed unusually high transient steam tem-peratures in the superheat range. This phenomenon oc-curred regardless of the internal pressure of the sterilizerchamber upon admission of steam, but was more pro-nounced with equipment employing an initial vacuum.In either case, saturated steam entering the sterilizervessel came into contact with the hot jacket and wassuperheated. This continued as long as the internal pres-sure of the chamber was lower than that of the jacket.As steam flow into the chamber continued, the super-heated steam in the chamber was compressed. This com-pression took place adiabatically for short periods oftime, resulting in steam temperatures far above the in-tended operating temperature of the device. In an emptysterilizer, the excess heat of the superheated steam dis-sipates to the walls of the vessel until the system returnsto the phase boundary. When a load of fabrics is presentin the chamber, however, the excess heat is given up to

the surface layers of the fabrics. Because of the insulatingproperties of dry fabrics and superheated steam (Henry,1959; Knox et al., 1960), the steam environment in theseareas did not return to the phase boundary for the dura-tion of the sterilizing cycle.The rate of spore destruction in regions of superheated

steam is less than in saturated steam at the same tem-perature. Under practical operating conditions, however,the temperature of the superheated steam is higher thanthat of the saturated steam in the system. In some casesthe increased temperature offsets small deviations fromthe phase boundary, and sterilization of spore controlscan be obtained with no increase in time. As deviationsfrom the phase boundary become large, they are notoffset by the increased temperatures, and total destructionof control spore populations does not result within thesame time period (Savage, 1937).The above considerations illustrate the necessity for

control of superheat in design and operation of steamsterilizers. Factors which affect the amount of super-heating and which can be controlled in design include thejacket temperature, initial chamber pressure, quality andrate of delivery of steam, and time elapsed between place-ment of materials in the chamber and input of steam.

ACKNOWLEDGMENT

The authors wish to thank Mildred Kandratavich fortechnical assistance with the bacteriological aspects of theinvestigation.

LITERATURE CITED

AMERICAN PUBLIC HEALTH ASSOCIATION, INC. 1953. Microbiologi-cal methods for milk and cream, p. 81-141. In Standardmethods for the examination of dairy products, 10th ed.American Public Health Association, New York.

BIGELOW, W. D. 1920. Heat penetration in processing cannedfoods. Natl. Canners Assoc. Research Lab. Bull. 16-L.

BOWIE, J. H. 1961. The control of heat sterilizers, p. 109-142. InSterilization of surgical materials. The PharmaceuticalPress, London.

FALLON, R. J. 1961. Factors concerned in the efficient steamsterilization of surgical dressings. J. Clin. Pathol. 14:505-511.

HALVORSON, H. 0. 1958. The physiology of the bacterial spore.Tech. Univ., Trondheim, Norway.

HENRY, P. S. H. 1959. Physical aspects of sterilizing cotton articlesby steam. J. Appl. Bacteriol. 22:159-173.

HUMPHREY, A. E., AND J. T. R. NICKERSON. 1961. Testing thermaldeath data for significant nonlogarithmic behavior. Appl.Microbiol. 9:282-286.

KELSEY, J. C. 1958. The testing of sterilizers. Lancet 1:306-309.KNOX, R., AND E. J. K. PENIKETT. 1958. Influence of initial

vacuum on steam sterilization of dressings. Brit. Med. J.1:680-682.

KNOX, R., E. J. K. PENIKETT, AND M. E. DUNCAN. 1960. Theavoidance of excessive superheating during steam sterilizingof dressings. J. Appl. Bacteriol. 23:21-27.

PENIKETT, E. J. K., T. W. ROWE, AND E. RoBSON. 1958. Vacuum

456 [VOL. 10

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/a

m o

n 19

Nov

embe

r 20

21 b

y 15

3.13

6.10

9.21

2.

1962] THERMAL DEATH OF SPORES IN STEAM 457

drying of steam sterilized dressings. J. Appl. Bacteriol. 21:282- apparatus for the multiple determination of rates of destruc-290. tion of bacteria and bacterial spores subjected to heat. Food

SAVAGE, R. M. 1937. Experiments on the sterilizing effects of Technol. 8:139-143.mixtures of air and steam, and of superheated steam. Quart. WALTER, C. W. 1948. Sterilization by steam, p. 54-73. In C. W.J. Pharm. and Pharmacol. 10:451-462. Walter [ed.], The aseptic treatment of wounds. The Mac-

STERN, J. A., AND B. E. PROCTOR. 1954. A micro-method and millan Co., New York.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/a

m o

n 19

Nov

embe

r 20

21 b

y 15

3.13

6.10

9.21

2.