bioreactor designs for solid state fermentation

its value expresses the capac-ity of the equipment to transfer oxygen independently of the volume of the reactor and so, constitutes an important parameter used for the scale-up studies

Is solid state fermentation (SSF) a new challenge for a

very ancient technology? The question is worth asking

be-cause this “ancient art” is in the process of becoming a

mod-ern technology

During the last 10 years, many articles have been

pub-lished, several books have been edited showing a spurt of

SSF processes even in western countries The fact that this

process is particularly well adapted to the metabolism of

fungi, the micro-organisms most commonly in SSF

pro-cesses, is an important feature because of the characteristics

of these micro-organisms (apical growth, enzymatic

ac-tivities) Moreover, in western countries, recent important

problems has emerged such as: pollution of soils and the

potential use of bioremediation, BSE epidemic and the

ne-cessity to find alternative for animal feeding, to cite only

two examples Thus, SSF has gained a new interest from

researchers and manufacturers over the past 10 years

Many papers have appeared on the use of solid state

fer-mentation, with studies on the effects of different factors on

fungus metabolism, and the potential for producing

differ-ent metabolites [1–3] The great majority of these papers

were SSF processes at laboratory-scale Conversely, very

few works have been carried out on the engineering aspects

and problems of scale-up

Compared to submerged fermentation, the solid media

used in SSF contain less water but an important gas phase

ex-ist between the particles This feature is of great importance

∗Tel.:+33-3-8069-3061; fax: +33-3-8069-3229.

E-mail address: durand@dijon.inra.fr (A Durand).

to the water Another point is the wide variety of matrices used in SSF which vary in terms of composition, size, me-chanical resistance, porosity and water holding capacity All these factors can affect the reactor design and the control strategy for the parameters Indeed in submerged fermenta-tion, we can consider roughly that all the media are made

up essentially of water In this environment, the temperature and pH regulations are trivial and pose no problem during the scaling-up of a process In submerged fermentation, only one major difficulty is encountered: the transfer of oxygen

to micro-organisms which depends upon the shape, the size

of the reactor and the agitation/aeration system used To

characterise this transfer, a parameter, KLa (oxygen trans-fer coefficient), has been defined It can be considered as

a “similarity invariant”, i.e its value expresses the capac-ity of the equipment to transfer oxygen independently of the volume of the reactor and so, constitutes an important parameter used for the scale-up studies in submerged fer-mentation In SSF, besides the oxygen transfer which can be

a limiting factor for some designs, the problems are more complex and affect the control of two important param-eters: the temperature and the water content of the solid medium

Other factors also affect the bioreactor design: (i) the mor-phology of the fungus (presence or not of septum in the hyphae) and, related to this, its resistance to mechanical ag-itation, (ii) the necessity or not to have a sterile process Before analysing the various types of bioreactors, their advantages and drawbacks, it is important to specify that

in a general way, many types of reactors are able to run at laboratory-scale with small quantities of medium But, the

1369-703X/02/$ – see front matter © 2002 Elsevier Science B.V All rights reserved.

PII: S 1 3 6 9 - 7 0 3 X ( 0 2 ) 0 0 1 2 4 - 9

scale-up is complicated mainly by intense heat generation

and heterogeneity in the system[4]

In this paper, emphasis will be put on the differences

between bench-scale bioreactors and pilot or industrial units

and also between non-sterile and sterile process

2 Bioreactor classification

Two categories of bioreactor exist for the SSF processes:

(i) at laboratory-scale, using quantities of dry solid medium

from a few grams up to few kilograms, (ii) at pilot and

industrial-scale, where several kilograms up to several tons

are used The first category comprises many designs, more or

less sophisticated, while the second category, which is used

mainly at industrial level, is markedly less varied Within

each category, some of the bioreactors can operate in aseptic

conditions

2.1 Laboratory-scale bioreactors

Several types of equipment are used for SSF Petri dishes,

jars, wide mouth Erlenmeyer flasks, Roux bottles and roller

bottles offer the advantage of simplicity [5,6] Without

forced aeration and agitation, only the temperature of the

room, where they are incubated, is regulated Easy to use

in large numbers, they are particularly well adapted for the

screening of substrates or micro-organisms in the first steps

of a research and development program

One of the interesting lab-scale units is the equipment

de-veloped and patented by an ORSTOM team between 1975

and 1980 [7] It is composed of small columns (Ø 4 cm,

length 20 cm) filled with a medium previously inoculated

and placed in a thermoregulated water-bath (Fig 1) Water

saturated air passes through each column This eqiupment is

widely used by many researchers and offers the possibility to

Fig 1 Typical lab-scale column reactor Several columns detailed on the right part of the figure are located in a water-bath for temperature control.

aerate the culture and also analyse the micro-organism res-piration by connecting the columns to a gas chromatograph with an automated sampler that routinely samples each col-umn This equipment is convenient for screening studies, op-timisation of the medium composition and measurement of

CO2produced The small quantity of medium (few grams) used and the geometry of the glass column is suitable for maintaining the temperature in the reactors (the heat removal through the wall seems to be sufficient) The design of this reactor, however, does not permit sampling during fermenta-tion and so it is necessary to sacrifice one entire column for each analysis during the process This equipment, with its advantages (forced aeration, cheap, relatively easy to use), can constitute a first step in the research

A new generation of small reactors was developed by an INRA-team in France a few years later The first model de-veloped [8] addressed problems concerning the regulation

of the water content of the medium A second model built during 2000 has been tested but has not been reported in the literature As shown in the photograph (Fig 2), this re-actor has a working volume of about 1 l Compared to the first model, the principal changes were the introduction of a relative humidity probe, a cooling coil on the air circuit and

a heating cover for the vessel These changes improved the regulation of the water content during the process As for the ORSTOM columns, the mini-reactors are filled with a medium previously inoculated in a sterile hood Each reactor

is automatically controlled by a computer Moreover, sam-ples can be taken by opening the cover in the presence of a flame without problem of contamination In this type of re-actor, the temperature and the water amount of the medium can be monitored by means of the regulation of the tempera-ture, relative humidity and flow rate of the air going through the substrate layer Different profiles for the air-inlet tem-perature and flow rate can be elaborated and generate useful information for the scaling-up studies

Fig 2 Photography and schematic of a lab-scale sterile reactor (1) Heating cover, (2) medium temperature probe, (3) stainless steel sieve, (4) air-inlet temperature probe, (5) relative humidity probe, (6) resistive heater, (7) water temperature probe, (8) massic flow meter, (9) level probe, (10) insulating jacket.

Fig 3 Rotating drum bioreactor (1) Air-inlet, (2) rotating joint, (3) coupling, (4) air nozzles, (5) air line, (6) rollers, (7) rotating drum, (8) solid medium, (9) rim.

Fig 4 Perforated drum bioreactor.

Another concept, based on continuous agitation of the

solid medium, was developed by several teams mentioned

below The bioreactors can be a rotating drum (Fig 3), a

per-forated drum (Fig 4) or an horizontal paddle mixer (Fig 5)

With or without a water-jacket, this type of reactor is

re-quired to be continuously mixed to increase the contact

be-tween the reactor wall and the solid medium and also to

provide oxygen to the micro-organism For rotating drum

bioreactors, as an horizontal cylinder, the mixing is

pro-vided by the tumbling motion of the solid medium which

may be aided by baffles on the inner wall of the rotating

drum (perforated or not) However, in all these reactors, the

mixing is less efficient than with a paddle mixer [9]

In-deed, agglomeration of substrate particles during the growth

of the mycelium can occur which increases the difficulty of

Fig 5 Photography of an horizontal paddle mixer used in the Wageningen University of Agriculture Schematic of a stirred horizontal bioreactor (1) Air-inlet, (2) temperature probes, (3) water-jacket, (4) paddles, (5) air outlet, (6) agitation motor, (7) reactor, (8) solid medium, (9) agitation shaft.

regulating the temperature of the solid medium Moreover, the oxygen transfer inside these balls of medium, agglomer-ated by the fungal hyphae and also very often by the stick-iness of the substrate used, may be very low or nil In ad-dition, from an engineering point of view, a water-jacket on

a moving body of a reactor causes problems that increase with scale[10]

A continuous mixing horizontal paddle mixer (Fig 5) was developed by a Dutch team at Wageningen University This aseptic fermenter was used for different purposes and to improve simultaneous control of temperature and moisture content Although heat transport to the bioreactor wall was improved, this device becomes inefficient for larger volume

[11] because heat removal only through the wall becomes increasingly inefficient as the volume increases

Generally, a continuous agitation, even if it is gentle, can modify the structure of the solid medium to a pasty texture Depending upon the nature of the particles (clay granules

as support for example), this agitation can also be abrasive and so be harmful for the mycelium especially if the hyphae have no septa

For processes in which the substrate bed must remain static, a reactor designed by ORSTOM team in France and named Zymotis is an interesting equipment[12,13] It con-sists of vertical internal heat transfer plates in which cold water circulates (Fig 6) Between each plate the previously inoculated solid medium is loaded Thermostated air is in-troduced through the bottom of each partition This reactor, which looks like a tray reactor where the layers of substrate would be set vertically, appears difficult to work in aseptic conditions

Very often in SSF a shrinkage of the volume of medium occurs during the mycelium growth With this type of device,

Fig 6 Photography of the Zymotis showing heat exchanger plates for the thermostated water circulation (at left) and during a culture (at right).

the risk is that the contacts with the vertical plates will

de-crease as the fermentation progresses, which would lead to

poor heat transfer and air channelling Finally, the scale-up

of such a design appears very difficult

2.2 Pilot and industrial-scale bioreactors

As mentioned before, the number of reactor types used at

pilot scale and in industry is less wide due, at once to some

important reasons and necessities which are that:

Fig 7 Koji-type reactor: (1) Koji room, (2) water valve, (3) UV tube, (4, 8, 13) air blowers, (5, 11) air filters, (6) air outlet, (7) humidifier, (9) heater, (10) air recirculation, (12) air-inlet, (14) trays, (15) tray holders.

• above some critical quantity of substrate, the heat removal

becomes difficult to solve and restricts the design strate-gies available The solid medium becomes compacted or creates air channelling, shrinkage, etc All these factors affect heat and mass transfer,

• the properties of the micro-organism with respect to its

resistance to mechanical stirring, its oxygen requirement and temperature range When the mycelium hyphae do not have septa, they can be destroyed by a mechanical stirring So, the culture layer will be thin to allow heat

removal which automatically orientates to a category of

reactor,

• the nature of the substrate and the need to pretreat or not

it, appropriate procedures for the inoculation, the sterility

or the level of contamination acceptable for the process

and the application,

• the economy of the country where the process is

devel-oped especially with respects to the labour cost Indeed

some technologies need more manpower than others,

• handling poses different problems such as the ease of

filling, emptying and cleaning the reactor

The heat and mass transfer problems identified above can

be attributed to poor aeration This problem can be addressed

using the following strategies: (i) the air circulates around

the substrate layer or (ii) it goes through it Within the

sec-ond strategy, three possibilities are available: unmixed,

in-termittently or continuously mixed beds

2.2.1 SSF bioreactors without forced aeration

This category is ancient and the simplest Probably

differ-ent ancidiffer-ent civilisations have used this technology

domesti-cally for fermenting miscellaneous raw agricultural products

in baskets The microbial starter culture might be transferred

in the form of a “mouldy medium” Although this technology

has advanced, it is still based on the same principle Applied

on commercial scale, it corresponds to the tray fermenters

(Fig 7) as typified by the famous Koji process[14–18] Made

of wood, metal or plastic, perforated or not, these trays,

con-taining the solid medium at a maximum depth of 15 cm, are

placed in thermostated rooms The trays are stacked in tiers,

one above the other with a gap of a few centimetre This

technology can be scaled-up easily because only the

num-ber of trays is increased Although it has been extensively

used in industry (mainly in Asian countries), this

technol-ogy requires large areas (incubation rooms) and is labour

intensive It is difficult to apply this technology to sterile

processes except if sterile rooms are built and if procedures

and equipment for the employees are provided, which will

be prohibitive An alternative could be to use polypropylene

semi permeable sterilizable bags to maintain sterility

More-over, some bags have a microporous zone which allows a

passive airflow rate from 20 to 2000 cm3/(cm2/min)

2.2.2 Unmixed SSF bioreactors with forced aeration

The basic design feature of packed-bed bioreactors is the

introduction of air through a sieve which supports the

sub-strate In this way, a bioreactor was developed at pre-pilot

scale (Fig 8) for defining the control strategy and

optimis-ing the air-inlet temperature, the airflow rate, the addition

of water and agitation during a SSF process [8] Located

in a clean room, the reactor can be pasteurised in situ by

steam generated by the water-bath used for the air

humidi-fication This reactor is very simple and can process a few

kilograms of dry solid medium These reactors constitute an

interesting tool that can be used in two ways: (i) to analyse

Fig 8 General view and schematic of the unmixed bioreactors with forced aeration (1) Basket conateining the solid medium, (2) valves for airflow adjustment, (3) air temperature probe, (4) relative humidity probe, (5) draincocks, (6) heating box, (7) humidifier, (8) coil for circulation of cold water, (9) resistive heater.

empirically the global evolution of a process and determine the environmental parameters for regulating the temperature and the moisture of the solid medium, (ii) to study mass and heat transfer phenomena and oxygen diffusion [19] Both the reactor diameter and the height of the substrate layer are around 40 cm, so the quantity of solid medium is suffi-cient to predict what can happen in a larger volume[20] In the absence of mathematical models for the scale-up, these reactors are very useful No mechanical agitation exists in-side these reactors, but the medium can be manually agi-tated in situ or it can be transferred into a kneading machine and reloaded into the basket However, this type of device without agitation is limited by the metabolic heat produc-tion Temperature gradients from the bottom to the top of

reactor was constructed by stacking and interconnecting

in-dividual modules (Fig 10) Non-communicating channels

Fig 9 Schematic of the patented industrial bioreactor showing the exchanger plates under each tray [22]

the headspace Different teams have worked on this design and it is mostly used at lab and pre-pilot scale Although

Fig 10 Scheamtic of the Plafractor TM reactor [23]

rotating drums have been described in the past, the largest

reactor recently cited in the literature was a 200 l stainless

steel rotating drum (Ø 56 cm and 90 cm long) which used

10 kg of steamed wheat bran as substrate[24] for kinetic

studies of Rhizopus Researches were carried out at lab-scale

to study the efficiency of this design, the role of the baffles

and the influence of the filling (amount of substrate per

unit volume) on the mass transfer by using tracer or image

analysis[25–27] These works introduced the rational design

and scale-up of this type of reactor In several cases, the

mycelium and the substrate particles, particularly starchy

and sticky materials, agglomerate Under these conditions,

even with baffles inside the drum, it was very difficult to

separate these aggregates, consequently, the heat, mass and

oxygen transfers were greatly reduced When the rotation

rate of the drum is increased, it can affect the mycelium growth presumably because of shear effects[25]

For a discontinuously rotating drum, the design is iden-tical to the reactor described above but between two agi-tations, it operates like a tray reactor So, it is absolutely necessary to limit the height of the substrate layer, other-wise it will be necessary to continuously agitate due to the heat accumulation and, taking into account the poor ther-mal conductivity of the air, the medium temperature will in-evitably increase Very few studies have been published on this type Using this rotating drum, a strategy for regulating the medium temperature was described in a thesis[28]and

in a publication [29] It consists of activating the rotation

of the drum in response to the temperature measured by a thermocouple in the medium Efficient for a 4.7 l working

Fig 11 Discontinuously rotating drum [28]

volume (Fig 11), on soy beans with Rhizopus, for tempe

production, no scale-up studies have been attempted

2.2.4 Intermittently mixed bed bioreactors with

forced aeration

In general, these bioreactors can be described as packed

beds in which conditioned air passes through the bed An

agitation device is periodically used to mix the bed and at

the same time, water is sprayed if necessary The design

of these reactors, the capacity of which varies from a few

Fig 12 Photography and schematic of the Koji making equipment: (1) Koji room, (2) rotating perforated table, (3) turning machine, (4, 11) screw and machine for unloading, (5) air conditioner, (6) fan, (7) air outlet, (S) dampers (9) air filter, (10) machine for filling, (12) control board.

kilograms to several tons, is influenced the necessity or not

to operate in sterile conditions

For non-sterile processes, a number of advances have

been done in the design and application of such bioreactors

One design is represented by the rotary type automatic Koji

making equipment marketed by Fujiwara in Japan (Fig 12) The treated substrate is heaped up on a rotary disc De-pending on the diameter of this disc, different working volumes are available but always with a layer of maximum thickness 50 cm This non-sterile reactor operates with a

microcomputer which controls all the parameters

(tempera-ture of the air-inlet, air flow rate and agitation periods) The

main drawback of this equipment is the need to prepare and

inoculate the substrate in other equipment before filling the

reactor Nevertheless this type of design is widely used in

Asian countries

Similar to the reactors used in the barley malting process,

huge equipment has been built for the first step of the process

for making soy sauce A specific building contains the solid

state reactor which is generally rectangular with a length

of several meters Several tons of pretreated and inoculated

substrate are put on a wire mesh and conditioned air is forced

through the layer An agitator trolley periodically mixes the

solid medium Although this kind of reactor is very simple

and basic, it is widely used in many Asian manufacturers of

soy sauces

An INRA team in Dijon (France) has developed a

non-sterile process strategy based on the following principle

(Fig 13) The temperature (Tm) and the moisture (WAm) of

medium are maintained by a regulation of the temperature,

relative humidity and flow rate of the air input It is also

nec-essary to spray water (E) and agitate (A) periodically The

Fig 14 Pilot plant reactor [30] : Photography showing a general view of the reactor (left), a detail of the swelling joints A schematic diagram of this pilot plant (1) Carriage motor, (2) screw motor, (3) valves for inoculum and water spraying, (4) temperature probes, (5) weight gauges, (6) relative humidity probe, (7) cooler, (8) humidifier by steam injection, (9) airflow meter, (10) fan, (11) heater, (12) air filter, (13) cooler.

Fig 13 General schematic of the intermittently packed bed reactor with

forced aeration (Tin, HRin and Din) respectively the temperature, relative humidity and flow rate of the air-inlet, (Tout, HRout and Dout), respectively, the temperature, relative humidity and flow rate of the air outlet, (Tm)

temperature of the solid medium, (WAm) water amount of the solid

medium, (Mg) total mass of the solid medium, (A) agitation, (E) water spray.