JOURNAL OF MATERIALS SCIENCE35 (2000 ) 261– 270

Review

Bacterial cellulose—a masterpiece

of nature’s arts

M. IGUCHI∗Bogor Research Station for Rubber Technology, Jl. Salak No.1, Bogor, 16151, IndonesiaE-mail: maiguch@ibm.net

S. YAMANAKAAjinomoto Central Research, 1-1 Suzuki-cho, Kawasaki, Kawasaki 201, Japan

A. BUDHIONOInstitute for Research and Development of Agro-Based Industry, Jl. Ir. Juanda No.11,Bogor 16122, Indonesia

Ever since its remarkable mechanical properties were found fifteen years ago, interest hasgrown in bacterial cellulose for which the use had been more or less limited to themanufacture of nata-de-coco, an indigenous food of South-East Asia. This paper reviewsthe progress of relevant studies including the production of cellulose by bacteria, theformation of microfibrils and gel layer, the properties of gel and processed sheets, andsome aspects of applications. C© 2000 Kluwer Academic Publishers

1. IntroductionWhilst “cellulose” is a word originally given, in earlylast century by Anselme Payen, to the substance whichconstitutes the cell wall of higher plants [1], bacterialcellulose is an ex-cellar product of vinegar bacteriawhich was described by Louis Pasteur as “a sort ofmoist skin, swollen, gelatinous and slippery. . .” [2].Although the solid portion in the gel-like stuff is lessthan one percent, it is almost pure cellulose containingno lignin and other foreign substances.

Most familiarly, bacterial cellulose has long beenuseful as the raw material ofnata-de-coco, an in-digenous dessert food of Philippines, for which one-centimetre thick gel sheets fermented with coconut-water are cut into cubes and immersed in sugar sirup.Similar food can be prepared from other saps or fruitjuices, e.g.,Nata-de-pinafrom pineapple. That the ma-jor component ofnata-de-cocogel was cellulose, notdextran as assumed in the past, was proved in 1960s [3].Nata-de-cocois now manufactured in a large quantityat the level of home industry also in Indonesia and ex-ported as a healthy diet.Teekvass, or tea-fungus, grownin tea-cups and served in some parts of Russia andMiddle-Asia is said to be a similar ferment [4].

Scientifically, a substance known as “vinegar plant”or “mother” and of use for vinegar brewery in old daysin Europe was cultured in pure condition and identifiedby Brown [5, 6] to be the same as cell-wall cellulosefrom its chemical composition and reactivity, althoughcontemporary means of microscopy only gave a picture

∗ To whom correspondence should be addressed.

of “bacteria lying embedded in a transparent structure-less film”.

With the emergence of X-ray diffraction early thiscentury, it was observed that bacterial cellulose be-longed crystallographically to Cellulose I, commonwith natural cellulose of vegetable origin, in which twocellobiose units were arranged parallel in a unit cell,and that cellulose molecules tended to have a specificplaner orientation in dried film [7]. The change of ori-entation in drying process was also studied in early days[8]. After the advent of electron microscope, the water-swollen cellulosic gel was revealed to comprise randomassembly of microfibrils of less than 100̊A diameter[9] such as seen in a scanning electron micrograph offreeze-dried gel surface in Fig. 1, whereas cell-wall

Figure 1 A scanning electron micrograph of freeze-dried surface of bac-terial cellulose gel.

0022–2461 C© 2000 Kluwer Academic Publishers 261

Figure 2 Schematic model of bacterial cellulose microfibrils (right)drawn in comparison with “fringed micelle” (left).

Figure 3 Chemical Abstract citation of bacterial-cellulose relatedarticles.

cellulose had much complicated structure. The struc-ture of microfibril should be as simple as drawn inFig. 2 compared to the “fringed micelle” envisaged inold days for the texture of vegetable fibres. Indeed, itis one of the finest examples of Nature’s arts in whichlong chain molecules are aligned parallel in the ex-tended form. One may recall that such an oriented struc-ture had been idealized in fibre spinning ever since theearly days of artificial silk manufacture, or at least af-ter the recognition of macromolecular hypothesis, andonly simulated in the last few decades in the efforts ofdeveloping super-strong fibres and growing extended-chain single-crystals.

According to Brown [5, 6], the pellicle of bacterialcellulose was “very tough, especially if an attempt was

made to tear it across its plane of growth”. There werea number of people who obtained films from the pel-licle and studied the structure, but somehow no atten-tion had been paid to the physical properties of filmsuntil mid-1980s when stress-strain measurement wasfirst conducted by the present authors [10–13]. TheYoung’s modulus recorded, 16–18 GPa isotropicallyacross the surface of plane, was extraordinarily large fortwo-dimensional materials of organic substances, andfurther improved up to 30 GPa. The fragments of bacte-rial cellulose were also found effective for reinforcingpulp papers and useful for other purposes. Among vari-ous possible applications, these materials have becomeof use for acoustic diaphragms of high-fidelity loud-speakers and headphones.

Interest in bacterial cellulose has grown rapidly inthe past decade as seen in the statistics of publicationsshown in Fig. 3. This paper is aimed at reviewing thedevelopment of bacterial cellulose study with specialreference to its use as materials.

2. Production of cellulose by bacteria2.1. Bacterial speciesThe species of bacteria which produces cellulose isgenerally calledAcetobacter xylinum, although bacte-ria of different names are often of use in literatures. Thevinegar-plant bacterium originally used by Brown wasobtained from a pellicle appeared on the surface of beer.In nature, the kind of bacteria are found, for instance,in rotten fruits and vegetables as more than thirty caseshave been reported [14]. The reason why the microor-ganisms generate cellulose has been a quest of biolo-gists. One considers that the aerobic bacteria producepellicle to maintain their position close to the surfaceof culture solution [15, 16]. Another assumes that thebacteria generate cellulose to guard themselves fromultraviolet [17]. The authors prefer to imagine that theyconstruct such a ‘cage’ and confine themselves in it toprotect themselves from enemies and heavy-metal ions,whereas nutrients can be supplied easily by diffusion.

2.2. Culture methodsThe source substance of bacterial cellulose is sac-charides. A typical culture medium widely of use inlaboratories is prepared by dissolving, 50 g sucrose, 5 gyeast-extract, 5 g (NH4)2SO4, 3 g KH2PO4, and 0.05 gMgSO4·7H2O in a litre of water [15]. According to theexperience of the authors, the recipe can be more com-plete if a small amount of vitamins is added. Althoughthe addition of inorganic nutrients are not necessarilyrequired when natural saps and juices are used, it isa common practice innata-de-cocoindustry to add asmall amount of nitrogen-containing compounds, suchas ammonium sulphate and di-ammonium hydrogenphosphate.

Culture is carried out normally in static condition ataround 28–30◦C by adding an aliquot of activated seedbroth to the culture medium. The system becomes tur-bid and, after a while, a white pellicle appears on thesurface and its thickness increases steadily with time,reaching over 25 mm in four weeks, as demonstrated inFig. 4. It is important to note that in the process of gel

262

Figure 4 Bacterial cellulose layers grown with different culture time(maximum 4 weeks).

growth the aerobic bacteria generate cellulose only inthe vicinity of surface, so that the productivity dependsprimarily on the surface area, not on the volume of ves-sel [18]. As long as the system is kept unshaken, thedisc-shaped gel is suspended by the cohesion to theinterior wall of vessel and slides steadily downwardsas it thickens. It was experienced by the authors thatthe growth of continuous gel layer tended to fail if avessel with tapered wall like a conical flask was used.In nata-de-cocohome industries, plastic vessels of ca.50w× 35d× 10h cm3 are employed. After the inocu-lation, the vessels are covered with an old newspaperand kept in a storehouse for 8–10 days. If purificationis necessary,viz. for scientific purpose, bacteria con-tained in the nascent gel can be conveniently removedby immersing it in dilute alkaline solution and wash-ing with water as originally conducted by Brown [5, 6].For further purification, treatment by oxidant was foundeffective as described below [12].

With the aim of enhancing the productivity, culturein agitated conditions has been studied recently in [19],

Figure 5 Schematic illustration of bacterial cellulose biogenesis and fibril formation [22].

although a flat gel is no longer obtained and the use hasto be limited to such applications as papermaking.

2.3. Formation of microfibrilsThe mechanism of formation as well as the struc-ture of microfibril has been studied extensively in re-cent decades combining the knowledge of biogenesis[20, 21]. Today, it is believed as illustrated in Fig. 5that cellulose molecules synthesized in the interior ofbacterial cell are spun out of ‘cellulose export compo-nents’ or nozzles to form a protofibril of ca. 2–4 nmdiameter, and the protofibrils are bundled in the formof a ribbon-shaped microfibril of ca. 80× 4 nm [22].

The kinetics of cellulose production by bacterium hasbeen studied since 1950s and it has been established thatthe yield of cellulose increases almost exponentiallywith time, at least in low conversion ranges, when cul-ture is carried out in agitated condition and sufficientoxygen is supplied from air. It is commonly assumedthat a bacterial produces a certain number of chain ini-tiators during its generation time to which monomerunits are added to form cellulose and that the popula-tion of bacteria obeys the law of bacterial growth. Thus,

Nt = N0eαt (1)

whereNt andN0 are the number of bacteria at timetand 0, respectively, and a constant,α is related to themean generation time of bacteria,τ by;

α =(

1

τ

)ln 2 (2)

A theory to express the yield and the degree ofpolymerization on account of the average lifetimesof bacteria and chain-growth was derived [23] andτ = 220–330 min was estimated from the data of yieldand average molecular-weight measurements. Similar

263

experiments were carried out andτ = 290–480 min and380–900 min were reported by other authors [2, 24].Regarding the lifetime of chain growth itself, evidenceshave been given by the same authors in that the aver-age molecular weight of cellulose continues to increaseover generation change, possibly up to several gen-erations or more, although contradictory results wereraised rather recently [25].

What happens to the fibril structure during the cell di-vision is a question. It was considered that the formationof a three-way branching was inevitable, if the extru-sion of fibril continued beyond generation from motherto two daughter cells [11]. The fibrils may be narrowerat the branching point, if not normal number of nozzlesare provided at the stage of cell division, but the recov-ery of normal diameter has to be a matter of time. Infact, fibrils on magnified photographs appear not nec-essarily and not always linear. The segmental lengthbetween branching points was estimated as 580–960µm [26] from the lifetime of bacterium [24, 27] andthe growth rate of fibril [20, 28], and 200–700µm fromthe counting of bacteria in the product. The existenceof such branches, if true, may relate to the toughness orthe resistance against stretching of gel sheets.

2.4. Formation of gel layerThe mechanism of gel formation was considered as fol-lows [11, 29]. In the initial stage, the bacteria increasetheir population by taking dissolved oxygen and pro-duce a certain amount of cellulose in the entire liquidphase as observed by the appearance of turbidity. Whenthe dissolved oxygen is used up, bacteria existing onlyin the vicinity of surface can maintain their activity toproduce cellulose. Although they may undergo cell di-vision, the population in the surface region does notincrease exponentially but should reach a certain equi-librium number, as excess others are occluded in thegel and brought into the depth. Those bacteria belowthe surface are not ‘dead’ but ‘asleep’, so that they canbe reactivated and used as the seed for new cultureoperation. Whether oxygen pressure higher than in airaccelerates the cellulose production is different matterand rather complicated [30].

Regarding the growth of gel layer in static condi-tion, it is a general trend observed [11, 18, 31] in thatthe thickness as well as the yield of cellulose increasessharply, after a few days of induction period, until therate tended to slow down after a week or ten days.Fig. 6 reproduces the results of recent experiment inwhich the base medium was coconut-water [32]. Thethickness, wet weight and dry weight followed simi-lar trend and the addition of sugar did not give muchdifference at least when it was above 1%. As saccha-rides, fructose which should have been generated bythe hydrolysis of sucrose was not detected due pos-sibly to the conversion to some other substances. Asshown in Fig. 7, the concentration of glucose did notnecessarily decrease monotonically, particularly whenthe concentration of added sugar was high, whereasthe concentration of sucrose decreased monotonicallytowards zero. It was considered that glucose was the

Figure 6 Changes of the thickness, wet weight and dry weight of gel withculture time. The base medium was coconut-water and 1% (NH4)2HPO4

and 1–4% sugar was added [32].

kind of saccharide which was digested by bacteria andconverted to cellulose.

Fig. 8 shows computer-simulated curves of glucoseconsumption, or cellulose production corresponding toFig. 7 (bottom), in the second stage of reaction obtainedon the following equations [32].

−∂Co

∂t= −Do∂

2Co

∂x2+ KCoCg (3)

−∂Cg

∂t= Dg∂

2Cg

∂x2+ KCoCg (4)

whereCo andCg are the concentrations of oxygen andglucose,Do and Dg are the diffusion coefficients of

264

Figure 7 Changes of glucose and sucrose concentrations with culturetime. The conditions are the same as in Fig. 6 [32].

Figure 8 A set of cellulose yield vs. time traces calculated on Equa-tions 1 and 2. (Cg,0= 0.6–3.0× 10–7 mol/µl, Do= 9.0× 10–12mm2/h, Dg= 4.0× 10–12 mm2/h, Co,0= 1.6× 10–8 mol/µl and K =1.5× 10–8µl/mol/h,1t = 1 h) [32].

oxygen and glucose, respectively,x is the depth fromthe surface, andK is an integrated rate constant of glu-cose consumption. There, mass-transfer by convectionwas neglected, and the effect of gel layer on diffusionwas not taken into account as the solid fraction wasless than one percent in volume. In a plot of glucoseand oxygen concentration against the depth from thesurface, it was recognized that glucose diffuses gradu-ally from the interior, whereas oxygen diffusing fromair did not penetrate deep consumed by the reaction.

3. Properties of bacterial cellulose3.1. Elastic properties of gelAlthough bacterial cellulose is obtained in the form ofa highly swollen gel, the texture is quite unique anddifferent from typical hydro-gels. Those readers whohave tastednata-de-cocoshould know that the originalelasticity would never recover once the gel is crashed.One may remember the flesh of squid, a typical ori-ental seafood, which hardly swell again after dried.These are ascribed to the fact that the elements whichconstitute the gel are microfibrils, not the segments ofchain molecules, such as in agar or gelatin gels, whichcan take thermodynamically stable form. Fig. 9 showsan example of compression stress relaxation curve inwhich the stress continued to fall beyond the period ofmeasurement [33]. With three-element Maxwell model,it was fit by:

f = 8.5 exp(−3.64× 10−3)+ 14.9

× exp(−7.00× 10−2)+ 45.2 exp(−5.40× 10−1)

(5)

Fig. 10 (top) and (bottom) show the changes of com-plex viscosity,η∗, storage modulus,G′ and loss mod-ulus G′′ measured by a parallel plate rheometer as afunction of strain and frequency, respectively [34]. Theresponse is linear up to strains of 5% and the fact thatG′ is significantly higher thanG′′ implies that the ma-terial has characteristics of rubber in the deformationrange. The complex viscosity decreased monotonicallybut the storage and loss moduli maintained a certainlevel against the increase of oscillation frequency.

Since the gel is hard to be stretched beyond severalpercent, efforts of orienting fibrils such as made in thepast [8] have been virtually in vain. Attempts of coldextrusion was not successful either [34]. A roll deviceto wind up thin gel, in the form of a continuous ribbon,from the surface of culture medium was invented andapplied, but the orientation observed by X-ray diffrac-tion was not necessarily high [35].

Figure 9 Compression stress relaxation of bacterial cellulose gel [33].

265

Figure 10 Complex viscosity,η∗, storage modulus,G′ and loss modulusG′′ measured by a parallel plate rheometer as a function of strain (top)and frequency (bottom) [34].

3.2. Mechanical properties of filmsTraditionally, films were prepared by drying a gel sheetin air on a flat surface, e.g., glass plate, by fixing thearea. In the newly developed heat-press method [10], itis important to place a gel sheet sandwiched betweenstainless-steel meshes and/or non-woven fabrics to fa-cilitate the escape of water. The results of tensile mea-surements of films prepared in various conditions are

TABLE I Mechanical properties of bacterial cellulose films prepared in various conditions

Culture Film Young’s Tensiletime Preparation Temperature Pressure thickness modulus strength Elongation(days) methoda (◦C) (kPa) (µm) (GPa) (MPa) (%)

7 Air-dry 20 0 — 16.9 256 1.77 Heat-press (⊥) 150 49 — 17.4 224 1.87 Heat-press (⊥) 150 49 — 18 231 1.87 Heat-press (⊥) 200 49 — 16.4 243 1.97 Heat-press (⊥) 150 49 — 16.9 260 2.17 Heat-press (⊥) 150 196 — 16.7 216 1.77 Heat-press (⊥) 150 490 — 17.5 155 1.47 Heat-press (⊥) 150 980 — 17 129 0.97 Heat-press (⊥) 150 1470 — 16.6 102 0.87 Heat-press (⊥) 150 1960 — 18.1 91 0.87 Heat-press (⊥) 150 49 — 16.1 221 1.97 Heat-press (‖) 150 49 — 15.9 205 1.85 Heat-press (⊥) 150 49 14 16.5 246 1.97 Heat-press (⊥) 150 49 37 16.1 217 1.714 Heat-press (⊥) 150 49 63 16.2 255 228 Heat-press (⊥) 150 49 159 15.1 199 1.7

a Press direction: (⊥); normal to the plane of growth, (‖); parallel for a cut-out strip.

Figure 11 Dynamic viscoelastic properties (E′ and tanδ) of typicalfilm measured as a function of temperature (top) and relative humidity(bottom) [33].

summarized in Table I. As far as the Young’s modulusis concerned, the values obtained were much the samewithout regard to the preparative condition, and the ten-sile strength as well as elongation tended to decreasewhen excess pressure was applied, due presumably tothe introduction of defects.

Fig. 11 (top) shows dynamic viscoelastic propertiesof typical film measured as a function of temperature[33]. Whilst the dynamic modulus,E′ decreased slowly

266

from 15 to 9 GPa with the increase of temperature,tan δ showed two maxima at around 50 and 230◦C,corresponding to the desorption of water and thedegradation of cellulose, respectively. The specimenshowed a typical water-sorption isotherm in which thewater regain at 100%RH was 9.3%. As a function ofrelative humidity, E′ decreased and tanδ increasedgradually as seen in Fig. 11 (bottom).

Figure 12 Scanning electron-micrographs of fracture edge of bacterial cellulose film [11].

Figure 13 Change of Young’s modulus by chemical treatment [12].

Morphologically, fibrils in the sheets appear to con-stitute a pile of thin layers, as seen in Fig. 12, regardlessto the press direction. This magnified picture remindsone of the structure of pulp papers in which hydrogen-bond between fibrils is believed to be the source ofstrength [36]. In the case of bacterial cellulose, thedensity of inter-fibrillar hydrogen-bonds must be muchhigher, as the diameter of fibrils is much smaller, and

267

this can be the reason why such a high Young’s modulusdevelops with this material.

Since traces of contaminants were suspected to affectthe formation of hydrogen-bond, treatment with oxi-dant and alkaline solutions have been attempted parallelwith careful chemical analysis [12]. Fig. 13 (left) and(right) show the change of the Young’s modulus withthe concentration of NaClO and NaOH, in which max-ima are found at around 0.5%, and 5%, respectively,before cellulose is damaged at higher concentrations.In terms of the Young’s modulus, the best result wasobtained when the material was soaked in 0.5% NaClOsolution in the stage of gel and treated with 5% NaOHsolution after processing into film. The Young’s modu-lus attained, 30 GPa isotropically across the film plane,is quite large compared to the theoretical value of cel-lulose along the chain direction, 173 GPa [37]. It is re-marked that the value is several times higher than thoseattained by synthetic polymers, e.g., two-dimensionallystretched polyester film.

3.3. Properties of sheets prepared withfragmented bacterial cellulose

Suspension of fragmented bacterial cellulose gel wasobtained by means of a bladed-blender. The Young’smodulus and tensile strength of composite sheets pre-pared by filtering the mixture of cotton lint pulp andfragmented bacterial cellulose is plotted against thefraction of the latter in Fig. 14, in which the reinforcingeffect is clear. The pure suspension gave a sheet likeparchment paper which measured a Young’s modulus,4.9 GPa.

More practical data [38] for applying bacterial cel-lulose to papermaking is shown in Fig. 15. While theincrease in Young’s modulus and tensile index is repro-duced in (top) and (middle), respectively, it is anotherinteresting effect that the folding endurance of pulp pa-pers can be significantly improved (bottom). The is due

Figure 14 Young’s modulus and tensile strength of sheets prepared fromthe mixture of cotton lint and fragmented bacterial cellulose (data from[11]).

Figure 15 Properties of papers prepared by mixing bacterial cellulose[38].

presumably to the peculiar property of fragmented bac-terial cellulose that it tends to stick on other substance[10, 11]. Fig. 16 shows a scanning-electronmicrographof glass-fibre on which bacterial cellulose fragmentsare entangled on the surface and binding the fibres.Thus, one can prepare self-supporting sheets fromnon-cellulosic fibres without adhesive by adding smallamount of disintegrated bacterial cellulose as shown inTable II.

268

Figure 16 Scanning electron-micrograph of glass-fibre on which bacte-rial cellulose fragments are entangled.

Figure 17 Specific Young’s modulus vs. internal-loss of various acous-tic materials.

Figure 18 Frequency characteristics of test speakers (16 cmφ cone-type full-range). Solid-line: bacterial cellulose composite, Dotted-line: conventionalpaper.

TABLE I I Breaking length of sheets from non-cellulosic fibres pre-pared by mixing disintegrated bacterial cellulose

Fibre BC Bk lengthMaterial (parts) (parts) (km)

Novoloid fibre (Kainolr KP0203) 95 5 0.33Novoloid fibre (Kainolr KP0203) 90 10 0.79Novoloid fibre (Kainolr KP0203) 80 20 1.67Carbon fibre (Toreca T008r 6mmL) 90 5 0.64Alumina fibre (Denka Arecen Bulkr) 90 5 0.24

4. ApplicationsAmong various applications studied so far, that whichhas reached the level of practical use is for acoustic di-aphragms as bacterial cellulose has been found to bearthe two essential properties, i.e., high sonic velocityand low dynamic loss (see, Fig. 17). In fact, the sonicvelocity of pure film was almost equivalent to those ofaluminium and titanium, while the tangent-delta was ina low range, 0.4–0.5. In the sound-pressure-level curvesof a composite-paper cone diagram, shown in Fig. 18,it is seen that both frequency response and second har-monic distortion are smoothed and extended to higherfrequency regions. Thus, hi-fidelity loudspeakers andheadphones have been marketed by Sony Corp.

The use of films as the raw material of conductivecarbon film was investigated and found excellent, al-though it still stays in the cradle of laboratory [39].

The use of fragmented bacterial cellulose for paper-making is promising and test pieces of flexure-durablepapers and high filler-content papers, ideal for bank-note papers and bible papers, have been prepared byMitsubishi Paper Mills Co. Fancy-papers with low-portion bacterial cellulose has been also prepared butit is not an application aimed at improving the physicalproperties.

Other ideas raised include the use of sheet or film asa temporary skin for medical care [40] and separationmembrane [41], the use of fragmented suspension asan viscosity enhancing agent for various purposes, etc.

269

5. ConclusionThe finding that the unique structure of bacterial cel-lulose offers interesting properties has taught us thatmore useful substances are left in Nature unknown tomankind. It is an expectation of authors that bacte-rial cellulose can contribute Indonesia and other low-latitude countries to promote high-tech industries basedon their indigenous materials. The cost of the materialestimated as about US$30/kg at dry-base is expected tobe lowered by the use of agricultural wastes as carbo-hydrate resources and the rationalization of productionprocess.

AcknowledgementsAmong many colleagues, thanks are due especially toDr. M. Mackley, University of Cambridge, and Prof. S.Yano, Nihon University, who generously permitted toinclude their unpublished results. This review is basedon a lecture given in “Green Polymer Week”, organ-ised by Dutch Polymer Institute and held in Eindhoven,Netherlands, 29 June–3rd July 1998. The authors thankProf. P. J. Lemstra, Eindhoven University of Technol-ogy, for his encouragement to write this article.

References1. T . P. N E V E L L andS. H. Z E R O N I A N, in “Cellulose Chem-

istry and Applications,” edited by T. P. Nevel and S. H. Zeronian(Howrwood Ltd., 1995) p. 15.

2. G. J. F. R I N G, in “Cellulose and Other Natural Polym. System,”edited by R. M. Brown (Plenum, New York, 1982) p. 299.

3. M . M . L A P A S, E. G. G A L L A R D O and M . A . P A L O,Philipp. J. Sci.96 (1967) 91.

4. G. C A M P B E L L-P L A T T , in “Fermented Foods of the World—ADictionary and Guide” (Butterworth, 1987).

5. A . J. B R O W N, J. Chem. Soc.49 (1986) 172.6. Idem., ibid. 49 (1986) 432.7. H. M A R K andG. V . S U S I C H, Z. Physikal. Chem. (B)4 (1929)

431.8. A . W. S I S S O N, J. Phys. Chem.40 (1936) 343.9. E. F R A N T S andE. S C H I E B O L D, J. Makromol. Chem.1 (1943)

4.10. M . I G U C H I , S. M I T S U H A S H I , K . I C H I M U R A , Y . N I S H I ,

M . U R Y U, S. Y A M A N A K A andK . W A T A N A B E , US Patent4,742,164 (1988).

11. S. Y A M A N A K A , K . W A T A N A B E , N. K I T A M U R A , M .I G U C H I , S. M I T S U H A S H I , Y . N I S H I and M . U R Y U,J. Mater. Sci.24 (1989) 3141.

12. Y . N I S H I , M . U R Y U, S. Y A M A N A K A , K . W A T A N A B E ,K . K I T A M U R A , M . I G U C H I andS. M I T S U H A S H I , ibid. 25(1990) 2997.

13. M . I G U C H I , S. Y A M A N A K A , K . W A T A N A B E , Y . N I S H I

andM . U R Y U, in Proc. Int. Meet. Polym. Sci. Technol, RolducPolym. Meet. 5, May 1990, edited by P. J. Lemstra and L. A.Kleintjens 44-50, London, 1991) p. 371.

14. E. G. J E S U S, R. M . A N D R E S andE. T. M A G N O, Philipp.J. Sci.100(1971) 41.

15. M . S C H R A M M andS. H E S T R I N, J. Gen. Microbiol.11 (1954)123.

16. S. V A L L A andJ. K J O S B A K K E N, ibid. 128(1982) 1401–1408.17. W. S. W I L L I A M andR. E. C A N N O N, Appl. Env. Microbiol.

55 (1989) 2448.18. S. M A S A O K A , T . O H E andK . S A K O T A, Ferment. Bioeng.

75 (1993) 18.19. F. Y O S H I N A G A, K . T O N O U C H I and K . W A T A N A B E ,

Biosci. Biotechnol. Biochem.61 (1997) 219.20. R. M . B R O W N J R., J. H. M . W I L L I S O N and C. L .

R I C H A R D S O N, Proc. Natl. Acad. Sci. USA73 (1976) 4565.21. K . Z A A R , Cytobiologie16 (1977) 1.22. R. M . B R O W N J R., in “Cellulose: Structural and Functional

Aspects” (Ellis Howoodide, 1989) p. 145.23. A . M . B R O W N, J. Polymer Sci.59 (1962) 155.24. M . M A R X -F I G I N I andB. G. P I O N, Biochimica Biophysica

Acta338(1974) 382.25. K . O K A J I M A , Y . M A T S U D A andK . K A M I D E , Polymer In-

ternational25 (1991) 145.26. K . W A T A N A B E , PhD desertation, Tokyo University, 1995.27. M . M A R X -F I G I N I andB. G. P I O N, Makromol. Chem.177

(1976) 1013.28. A . K A I , J. K O G U S U R I andY . K O B A Y A S H I , Bull. Chem.

Soc. Jpn.(1982) 1394.29. M . I G U C H I andS. Y A M A N A K A , (1997) in Proc. Int. Work-

shop Green Polym., Bandung-Bogor 1996, edited by H. Adisessa(Indonesian Polymer Association, 1997) p. 47.

30. K . W A T A N A B E and S. Y A M A N A K A , Biosci. Biochem.59(1995) 65.

31. W. B O R Z A N I and S. J. D E S O U Z A, Biotechnol. Lett.17(1995) 1271.

32. A . B U D H I O N O, B. R O S I D I, H . T A H E R andM . I G U C H I ,Carbohydrate Polymers, in press.

33. S. Y A N O , A . B U D H I O N O andA . S U G I H A R T O, Report onthe International Research Cooperation Project: The industrial uti-lization of bacterial cellulose, AIST, Tokyo, 1998.

34. T . E. B U S T N E S, R. J. M U R R A Y and M . M A C K L E Y ,Project Report, Department of Chemical Engineering, Universityof Cambridge, 1998.

35. N. S A K A I R I , H . A S A N O, M . O G A W A, N. N I S H I andS.T O K U R A, Carbohydr. Polym.35 (1998) 233.

36. D. H. P A G E andR. S. S E T H, Tappi63 (1980) 113–116.37. K . T A S H I R O andM . K O B A Y A S H I , Polymer Bull. (Berlin)14

(1985) 213.38. H. S H I B A Z A K I , S. K U G A andF. O N A B E, Tappi, Jpn.48

(1994) 93.39. K . Y O S H I N O, R. M A T S U O K A , K . N O G A M I , H. A R A I ,

S. Y A M A N A K A , K . W A T A N A B E , M . T A K A H A S H I andM . H O N D A, Synthetic Metals41–43(1991) 1593.

40. J. D. F O N T A N A, A . M . D E S O U Z A, C. K . F O N T A N A,I . L . T O R R I A N I , J. C. M O R E S C H I, B . J. G A L L O T T I ,S. J. D E S O U Z A, G. P. N A R C I S C O, J. A . B I C H A R A andL . F. X . F A R A H, Appl. Biochem. Biotechnol.24–25(1990) 253.

41. H. S H I B A Z A K I , S. K U G A , F. O N A B E and M . U S U D A,J. Appl. Polym. Sci.50 (1993) 965.

Received 20 Apriland accepted 18 May 1999

270