APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Jan. 1986, p. 39-430099-2240/86/010039-05$02.00/0Copyright X 1986, American Society for Microbiology

Effect of Sterilization by Dry Heat or Autoclaving on BacterialPenetration through Berea Sandstone

GARY E. JENNEMAN,1 MICHAEL J. McINERNEY,l* MICHAEL E. CROCKER,2 AND ROY M. KNAPP3

Department ofBotany and Microbiologyl and School ofPetroleum and Geological Engineering,3 University of Oklahoma,Norman, Oklahoma 73019, and National Institute for Petroleum and Energy Research,2 Bartlesville, Oklahoma 74005

Received 1 July 1985/Accepted 16 October 1985

A study was undertaken to determine why bacteria could penetrate lengths of consolidated sandstone (Berea)faster when the sandstone was sterilized by autoclaving than when dry heat (150°C, 3 h) was used. Changes inpermeability, porosity, and pore entrance size of the rock as a result of autoclaving were not sufficient to explainthe differences in penetration times observed, but electron dispersion spectroscopy and electron microscopy ofthe rock revealed changes in mineral composition and clay morphology. Autoclaved cores contained more

chloride than dry-heated cores, and the clays of autoclaved cores were aggregated and irregularly shaped.Therefore, the decreases in bacterial penetration rates caused by autoclave sterilization were probably theresult of a change in surface charge of the pores of the rock and of a reduction in surface area of clays availablefor adhesion. The results implied that dry-heat sterilization was preferable to autoclaving when examiningbiotic and abiotic interactions in a native-state rock model.

Recent interest in subsurface environments with regard tometabolism of toxic wastes (8, 12), biofilm formation (9), andmicrobially enhanced oil recovery processes (3, 5) hascreated a need to develop experimental models to study thefactors influencing the rate of microbial penetration throughporous media. Some investigators have studied these factorsin synthetic soil or rock systems by packing glass columnswith sand, fused glass beads, iron filings, or even agar (13).These systems are useful and advantageous because of theease of preparation, consistency between columns, and thelimited number of abiotic variables. However, these exper-imental systems often do not adequately emulate the physi-cal environment of deep subsurface environments. This isthe reason that much of the current research on microbiallyenhanced oil recovery has been done with Berea sandstone.This rock is a consolidated sandstone outcropping which isquarried near Cleveland, Ohio. It is widely used in thepetroleum industry as a model rock system for studyingvarious oil recovery techniques because of its ready avail-ability and its close resemblance in many respects to thephysical properties of actual oil-bearing sandstones.To study penetration of selected bacterial species in

nutrient-saturated Berea sandstone it was necessary to ster-ilize the cores before inoculating them with the bacteriumunder investigation. In previous studies autoclaving hasmost often been used to sterilize sandstone cores (1, 6, 7).However, dry heat (i.e., 150°C for 3 h) is also an effectivemeans of sterilization. The purpose of this study was toexamine the effects of dry-heat sterilization and autoclavingon bacterial penetration times and to determine the physicalchanges that occurred in the Berea sandstone as a result ofthese treatments. It is hoped that by understanding thesechanges, insight may be gained as to ways to increasepenetration rates and to improve sterilization methods forconsolidated rock experimental models. Understanding theinteraction between biotic and abiotic factors governingbacterial penetration through porous media is paramount tothe development of microbially enhanced oil recovery andother environmental bioengineering processes.

* Corresponding author.

MATERIALS AND METHODS

Organisms and cultivation medium. Strain 140 was isolatedfrom a subterranean sample underlying the NormanMunicipal Landfill. BCI-1NS was isolated from a Bereasandstone core (Cleveland Quarries, Amherst, Ohio). Bothisolates are morphologically similar to cells of the genusBacillus (7).The growth medium (medium E with 0.1% [wt/vol]

NaNO3 and 0.05% [wt/vol] yeast extract) and cultivationconditions for both strains have been described previously(7).Core preparation. For experiments involving cell penetra-

tion, cores were prepared as described previously (7). Thesides of each core were coated with epoxy. Berea sandstonefrom Cleveland Quarries was used in all experiments. Thepermeability of 3/4-in.-diameter (1.91 cm) cores was deter-mined by using a permeameter and Darcy's law as describedpreviously (7); the results are given in millidarcys.

Sterilization of cores. (i) Autoclave. Cores used in penetra-tion studies were presaturated under vacuum with 5%(wt/vol) NaCl solution. The saturated cores were submersedin the brine and autoclaved at 121°C and 15 lb/in2 for 30 min.The initial permeability (K1) was determined by using 5%NaCl solution previously filtered through a 0.22-,um mem-brane filter (Millipore Corp., Bedford, Mass.). The coreswere then placed in growth chambers containing the growthmedium as described previously (7) and autoclaved for anadditional 20 min.

(ii) Dry heat. Cores were sterilized by dry heat in aforced-draft oven (Allied Fisher Scientific Co., Dallas, Tex.)at 150°C. The cores were saturated with the 5% NaClsolution and the Ki was determined. The cores were thenplaced in a glass petri dish and heated. Dry-heat-sterilizedcores were removed aseptically from the petri dish andvacuum saturated with sterile medium E containing 0.1%(wt/vol) NaNO3 and 0.05% (wt/vol) yeast extract. They werethen aseptically placed in growth chambers.

Penetration time. Penetration time was the time (in hours)elapsed from inoculation of one flask of the growth chamberuntil faint visible turbidity occurred in the other flask of thegrowth chamber (7).

39

Vol. 51, No. 1

on June 16, 2019 by guesthttp://aem

.asm.org/

Dow

nloaded from

40 JENNEMAN ET AL.

TABLE 1. Effect of autoclaving or dry-heat sterilization ofsandstone cores on the rate of microbial penetration through

Berea sandstone coresa

Sterilization procedure

Strainb Core Dry heatc AutoclavedCno. Penetra- K, Kf Penetra- K1 Kf

tine(h) (mdarcy) (mdarcy) tine(h) (mdarcy) (mdarcy)

140d 132 103 213.8 NDe ND ND ND133 103 314.5 ND ND ND ND134 ND 245.9 323.0 ND ND ND

BCI-1NS 132 96 365.6 ND 16 399.2 ND133 60-75 445.3 ND 18 460.0 ND134 ND 429.3 409.3 ND 421.6 510.2

a Each core was 7.4 cm long and 1.9 cm in diameter. The same cores wereused throughout the experiment, with core 134 serving as an uninoculatedcontrol. The sequence in which the treatments were applied was as follows:dry-heat sterilization, inoculated with strain 140; dry-heat sterilization, inocu-lated with BCI-1NS; and autoclaved, inoculated with BCI-1NS.

b Both strains are motile and were grown at 50°C in medium E plus 0.1%NaNO3 and 0.05% yeast extract.

c Dry-heat sterilization was at 150°C for 3.0 h. Autoclaving was for 20 min at121'C. K1 and Kf refer to core permeabilities measured before inoculation andafter incubation, respectively.

d It was later found that strain 140 was capable of penetrating in 27.5 h a 6.1-cm-long Berea core (K,, 202.0 mdarcys) that had been sterilized by autoclav-ing.

' ND, Not determined.

Pore entrance size distribution. The centrifugation methodof Slobod et al. (10) as modified by Torbati et al. (11) wasused to determine the capillary pressure profiles of thesandstone cores at increasing centrifugal forces. An ultra-centrifuge (model L5-50B; Beckman Instruments, Inc.,Fullerton, Calif.) was used. Pore entrance size distributionwas generated from the capillary pressure curve and fromphysical data on the length, diameter, dry weight, and wetweight of the core by using a computer program (E. C.Donaldson, personal communication).EDS and SEM. Both electron dispersion spectroscopy

(EDS) and scanning electron microscopy (SEM) of sand-stone were done as previously described by Crocker et al.(4). Three Berea cores (2- by 6-in. ([5- by 15-cm] cylinders)were cut from a 12- by 12- by 6-in. (30- by 30- by 15-cm)block of sandstone and steam cleaned for 2 weeks to removehumic substances as described previously (F. D. Sutterfield,M. S. thesis, University of Tulsa, Tulsa, Okla., 1973). Eachcore was saturated with 5% (wt/vol) NaCl solution undervacuum. Core 1 was sterilized by autoclaving at 121°C and15 lb/in2 for 1 h while submerged in brine. Core 2 wassterilized by heating in a forced-draft oven at 150°C for 3 h.Core 3 was steam cleaned but not brine saturated or steril-ized. Samples for analysis were taken from both ends of eachcore and from a point midway between the two ends.

RESULTS AND DISCUSSIONThe penetration time for strains BCI-1NS and 140 was

dependent on the method of core sterilization used (i.e., dryheated or autoclaved). The same cores (132, 133, 134) wereused for all penetration experiments shown in Table 1 toeliminate variations between different cores. Cores that weresterilized by dry heat had longer penetration times comparedwith autoclaved cores of the same length. Both sterilizationmethods increased the permeability of inoculated anduninoculated cores which indicated that changes in perme-

ability cannot explain the differences in penetration timesobserved with the two sterilization methods. The times forthe penetration of 7.4 cm of dry-heat-sterilized rock were 103h for both cores inoculated with strain 140, whereas, strainBCI-1NS took from 60 to 75 or 96 h to penetrate the samecores after additional dry-heat sterilization. When thesesame cores were then autoclaved, the time for penetrationby BCI-1NS was four to six times faster. These penetrationtimes for autoclaved cores are well within the times previ-ously reported for strain BCI-1NS under similar conditions(7).

Successive autoclaving decreased the penetration times ofstrain BCI-1NS through Berea cores of 7.4 to 7.6 or 11.2 to11.5 cm in length (Table 2). The penetration times of 7.4- to7.6-cm-long cores in Table 2 are similar to the penetrationtimes reported in Table 1 after autoclaving. The 3.6- to3.7-cm-long cores had faster penetration times that wereunaffected by successive autoclaving.

Strain 140 penetrated a 6.1-cm-long autoclaved Berea core(K1, 202.0 mdarcys) in 27.5 h or approximately 3.5 timesfaster than it took to penetrate a 7.4-cm-long core that wasdry-heat-sterilized. Jenneman et al. (7) previously showedthat penetration rate was independent of length for coreswith permeabilities above 200 mdarcys by using autoclavedBerea cores and strain BCI-1NS. So the small difference inlength (6.1 versus 7.4 cm) cannot account for the largeincrease in penetration time observed when strain 140 wasgrown in an autoclaved core. The presence of either a

dry-heat-sterilized core or an autoclaved core did not inhibitgrowth of BCI-1NS in liquid culture (results not shown).Therefore, it was concluded that some change occurred inthe sandstone during autoclaving.The modest increases in permeability observed after

autoclaving previously dry-heat-sterilized cores (Table 1) didnot seem sufficient to account for the large decreases ob-served in penetration times because the penetration rate ofBCI-1NS in Berea cores with permneabilities above 200mdarcys is not affected by increasing permeability (7).Therefore, it was necessary to examine what other physicalchanges in the rock occurred as a result of autoclaving anddry-heat sterilization.The autoclaving process might have increased the sizes of

pore entrance in the rock, thereby allowing the cells easier

TABLE 2. Effect of successive autoclave treatments onpenetration time of BCI-1NS through Berea sandstone cores

Penetration time afterCore no. K*a indicated autoclaveand length (mdarcy) treatment (h)b

1 2 3

B60, 3.6 269 8.5 9.0 10.0B61, 3.7 205 9.5 11.0 11.0B62, 3.6 405 10.5 11.0 11.0B57, 7.5 521 21.0 16.0 10.5B58, 7.4 478 17.5 16.0B59, 7.6 484 21.0 13.5 11.5B63, 11.4 379 37.5 23.0 18.0B64, 11.5 423 24.0 18.5B65, 11.2 367 29.5 27.0 18.5

a Permeabilities were measured after the initial 30-min autoclaving period(see Materials and Methods).

b Each core was first autoclaved for 30 min, flushed with medium E with0.05% yeast extract and 0.1% NaNO3, mounted in the growth chamber, andinoculated with BCI-1NS. After growth was observed in flask B, the treat-ment was repeated, except the core was autoclaved for 20 min beforemounting and inoculating.

APPL. ENVIRON. MICROBIOL.

on June 16, 2019 by guesthttp://aem

.asm.org/

Dow

nloaded from

BACTERIAL PENETRATION THROUGH SANDSTONE 41

Ii0 E~

0-

5-

0-0

Pore Throat Size (jpm)60^ B

55- 30 Untreated

50- 0 Dry heated

45- A Autocloved

40-

35

5

5-

rii~~~~~~~~~~~~~~~~~~~IO I I3 4o10 20 30 40 60 70 80 90

Pore Throat Size (jam)FIG. 1. Effect of autoclaving and dry heat on pore entrance size

distribution of Berea sandstone. The pore entrance size distributionof four cores cut from a block of Berea sandstone (400 mdarcys) was

determined before heating. (A) Two cores were autoclaved at 1210Cfor 30 min and then dry heated for 3 h at 150°C. (B) The other twocores were dry heated and then autoclaved as described in panel A.Pore entrance size distribution was determined after each heatingstep. Data obtained from only one of the replicate cores is shownbecause similar results were obtained for each core.

access through the core. However, the results from poreentrance size determinations of four different cores indicatedthat neither autoclaving nor dry-heat sterilization altered the

TABLE 3. Effect of dry heat and autoclaving on the static andnonstatic porosity of Berea sandstone cores

Porosity (mdarcy) by treatment:Core Static' Nonstatic'length(cm) Un- Dry Auto- Un- Dry Autoclavedc

treated heatc clavedc treated heatc Test 1 Test 2

3.7 0.214 0.212 0.2123.6 0.214 0.205 0.2043.6 0.214 0.211 0.2082.8 0.227 0.220 0.2207.2 0.209 0.201 0.218 0.2177.4 0.204 0.200 0.211 0.2117.3 0.203 0.200 0.216 0.214

a Cores used for static porosity measurements were not epoxied and were2.54 cm in diameter. All cores were cut from a block of Berea standstone (400+ 50 mdarcy).

b Cores used for nonstatic porosity were epoxied on the sides and were 1.90cm in diameter. Flow through cores was performed after each treatment at adifferential pressure of 1.02 atm (103.32 kPa). Porosities were determined afterflowing brine (5% NaCI) through core.

c Dry-heat treatment was for 3 h at 150°C, and autoclaving was for 20 min at121°C while the core was submerged in 5% NaCI brine.

distribution of pore entrance sizes in the sandstone (Fig. 1).Furthermore, the order of treatment (autoclave then dry heat[Fig. 1A] or dry heat then autoclave [Fig. 1B]) had littleeffect on the pore entrance size distribution. However, therewere differences in distribution of pore entrance sizes be-tween the different cores which is indicative of the inherentheterogeneity of consolidated rock samples.Another physical property of the rock that might have

been affected was porosity. Porosity is a measure of thefraction of void space to bulk volume in a porous medium.The effect of autoclaving and dry-heat sterilization on staticand nonstatic porosity is shown in Table 3. Nonstaticporosity refers to the porosity determined after 5% (wt/vol)NaCl solution was flushed through the core under a differ-ential pressure of about 2.5 atm (253.2 kPa), whereas staticporosity was measured without fluid flow through the core.Dry-heat sterilization did not appear to significantly changethe static or nonstatic porosity of the cores tested, nor didautoclaving have a significant effect on the static porosity ofthese same cores. However, the nonstatic porosity ofautoclaved cores increased by an average of 7.3% abovevalues for the dry-heat-sterilized cores (Table 3). However,the increase in nonstatic porosity of autoclaved cores wasnot statistically significant at P < 0.1 when a two-tailed t testfor the difference between two means was used (results notshown). Additional autoclaving had little or no effect onsubsequent nonstatic porosity measurements (Table 3).Crocker et al. (4) showed that the clay minerals in Berea

TABLE 4. Effect of dry heat and autoclaving on elemental composition of the pore surface of Berea sandstone as determined by EDS% Oxide

TreatmentaSiO2 Al,03 Fe2O3b CaOb TiO2b K20 Na2O ClO,

None 88.7 + 2.1c 8.7 ± 1.4 0 1.0 0 2.3 ± 0.3 0 NDdDry heat 84.9 ± 0.2 8.9 ± 0.4 0 0.3 0 2.2 ± 0.2 0 4.0 ± 0.4Autoclave 72.8 ± 10.9 3.6 ± 1.6 0 0 0 0.25 ± 0.4 0 23.5 ± 12.6

a Each core was prepared as described in the Materials and Methods section.b Values obtained from only one or two subsamples taken along the length of the core.C Mean ± standard deviation of the values of the three subsamples taken along the length of the core.d ND, Not determined.

O.- 3'!0~

U)

L-

U 1a)

0~

0

VOL. 51, 1986

1-1u

on June 16, 2019 by guesthttp://aem

.asm.org/

Dow

nloaded from

42 JENNEMAN ET AL.

FIG. 2. SEM of clay particles (arrows) from cores which were

not heated (A), dry heated (B), or autoclaved (C). Note the more

rounded edges and aggregation of clays in the autoclaved core. Bar= 10,m.

sandstone act to cement together the silica grains lining thepores as opposed to other sandstones in which the clays aredispersed freely within the pores. Therefore, it is possiblethat heating the cores in the presence of a hydrating phasecould cause a dissolution of the clay and silica particleswhich would result in a change in the degree of cementationand hence porosity. This effect would be especially pro-nounced if the cores were subjected to a subsequent brine

flush as was done when measuring nonstatic porosity. Bak-ing sandstone cores (usually at 600°C) tends to stabilize claysand fines and reduces their movement in the presence of aflowing fluid (4). This may explain why nonstatic porositydid not increase with dry-heated cores.EDS analysis of the surface of the dry-heat-sterilized

cores and the autoclaved cores revealed some interestingdifferences (Table 4). The surfaces of the nontreated anddry-heated Berea sandstone cores had higher silica, alumi-num, and potassium contents than the autoclaved cores. Thedry-heated core had a much lower chloride content than theautoclaved core. Other ions (Fe3+, Ca2+, and Ti4+) made uponly a small or negligible part of the surface components. Inall cases, the total amount of elemental oxides detected wasnear 100% of the total oxide measured in the sample so thatif other ions were present they would be in very smallquantities. These data are consistent with a change inporosity after autoclaving because a reduction in silicacontent indicates a loss in cementation between clays andthe grain particles (4). Furthermore, decreases in theamounts of aluminum and potassium in autoclaved coresamples are also consistent with a loss in clay content sinceclays are composed of these elements.More importantly, these analyses suggest that a change in

surface charge resulted from autoclaving. The large increasein chloride content and the reduction in the content ofaluminum and potassium could have shifted the surface to amore negatively charged state which could cause a greaterelectrostatic repulsion between negatively charged bacterialcells and mineral surfaces. Chang and Yen (2) found that theaddition of the anion pyrophosphate to cell suspensionsdecreased the degree of cell retention inside Berea sand-stone. They attributed this to a higher repulsive energyinduced at the surface of either the bacterial cell or theparticle surface, thereby reducing the possibility of irrevers-ible adhesion of cells.SEM of untreated, dry-heated, and autoclaved Berea

cores also showed that autoclaving changed the nature of theclay particles inside the core (Fig. 2). The untreated anddry-heated cores contained clay particles that were flat andplatelike with irregularly shaped edges. The clay particles inthe autoclaved core were clumped, with smooth androunded edges.The effect of autoclaving is manifold. Increases in the

porosity of the autoclaved cores along with decreases in theamount of silica, aluminum, and potassium, which are ele-ments often associated with clays, imply that autoclavingcauses the removal of clay particles. However, no changeswere observed in pore entrance size distribution, whichwould argue that these losses in cementation are not signif-icant with respect to pore geometry. On the other hand, EDSof the surface of the autoclaved core showed a large increasein the chloride content compared with dry-heat-sterilizedand nontreated cores. This could result in an increasednegative charge at the surface of the rock which would causea greater electrostatic repulsion between the negativelycharged surfaces of the bacterial cells and the rock. SEM,although it does not indicate any quantitative change in claycontent, shows that the clays have been structurally alteredfrom their natural sharp-edged appearance to a morerounded edge and aggregated arrangement. These aggre-gated particles would have less surface area available foradhesion of cells than the flat, platelike clay particles ob-served in the nontreated and dry-heated samples. Therefore,it is likely that autoclaving resulted in an alteration of claysand of the surfaces lining the pores. In turn, these alterations

APPL. ENVIRON. MICROBIOL.

on June 16, 2019 by guesthttp://aem

.asm.org/

Dow

nloaded from

BACTERIAL PENETRATION THROUGH SANDSTONE 43

probably resulted in an increase in the content of negativelycharged ions at the surfaces and in a reduction in surfacearea available for cell adhesion.Understanding the effect of sterilization on the physical

properties of porous rock and other porous media is crucialin explaining the role of biotic and abiotic factors in cellularpenetration or retention. The information provided by thiswork should prove useful in future comparisons of bacterialpenetration times and rates in different porous media. Inaddition, the results suggest that dry-heat sterilization maybe a better method than autoclaving to determine the pene-tration rate of bacteria in native-state cores. This informa-tion will be useful for future studies of processes involvingthe cleanup of toxic wastes polluting groundwater and ofmicrobially enhanced oil recovery, both of which mayrequire transport of injected microorganisms through con-solidated rock.

ACKNOWLEDGMENTSThis work was supported by Department of Energy contract no.

DE-AC19-80BC10300 and by the Energy Resources Institute of theUniversity of Oklahoma.We thank H. M. Torbati and E. C. Donaldson for assistance in

determining pore throat size distribution and D. E. Menzie andE. C. Donaldson for helpful comments.

LITERATURE CITED1. Chang, P. L., and T. F. Yen. 1984. Interaction of Escherichia

coli B and B/4 and bacteriophage T4D with Berea sandstonerock in relation to enhanced oil recovery. Appl. Environ.Microbiol. 47:544-550.

2. Chang, Y., and T. F. Yen. 1985. Effects of sodium pyrophos-phate additive on the "huff and puff" nutrient flooding MEORprocess, p. 247-256. In J. E. Zajic and E. C. Donaldson (ed.),Microbes and oil recovery, vol. 1. Bioresources Publishing, ElPaso, Tex.

3. Clark, J. B., D. M. Munnecke, and G. E. Jenneman. 1981. Insitu microbial enhancement of oil recovery. Dev. Ind. Micro-biol. 22:695-701.

4. Crocker, M. E., E. C. Donaldson, and L. M. Marchin. 1983.Comparison of reservoir rocks and related clays. Proceedings ofthe 58th Annual Technical Conference and Exhibition, p. 1-5.Society of Petroleum Engineers of AIME, San Francisco, Calif.(Preprint 11973.)

5. Finnerty, W. R., and M. E. Singer. 1983. Microbial enhance-ment of oil recovery. Bio/technology 1:47-54.

6. Jang, L.-K., P. W. Chang, J. E. Findley, and T. F. Yen. 1983.Selection of bacteria with favorable transport propertiesthrough porous rock for the application of microbial-enhancedoil recovery. Appl. Environ. Microbiol. 46:1066-1072.

7. Jenneman, G. E., M. J. McInerney, and R. M. Knapp. 1985.Microbial penetration through nutrient-saturated Berea sand-stone. Appl. Environ. Microbiol. 50:383-391.

8. McCormick, D. K. 1985. One bug's meat.... Bio/technology3:429-436.

9. Shaw, J. C., B. Bramhill, N. C. Wardlaw, and J. W. Costerton.1985. Bacterial fouling in a model core system. Appl. Environ.Microbiol. 49:693-701.

10. Slobod, R. L., A. Chambers, and W. L. Prehin, Jr. 1951. Use ofcentrifuge for determining connate water, residual oil and cap-illary pressure curves of small core samples. Petroleum Trans-action, AIME 192:129-134.

11. Torbati, H. M., E. C. Donaldson, G. E. Jenneman, R. M.Knapp, M. J. McInerney, and D. E. Menzie. 1985. Depth ofmicrobial plugging and its effect on pore size distribution inBerea sandstone cores, p. 213-225. In J. E. Zajic and E. C.Donaldson (ed.), Microbes and oil recovery, vol. 1. Interna-tional Bioresources Journal, Bioresources Publishing, El Paso,Tex.

12. Wilson, J. T., J. F. McNabb, D. L. Balkwili, and W. C. Ghiorse.1983. Biotransformation of selected organic pollutants ingroundwater. Dev. Ind. Microbiol. 24:225-233.

13. Wimpenny, J. W. T. 1981. Spatial order in microbial ecosys-tems. Biol. Rev. 56:295-342.

VOL. 51, 1986

on June 16, 2019 by guesthttp://aem

.asm.org/

Dow

nloaded from