You are on page 1of 22

Journal of Microbiological Methods 10 (1989) 155-176


Elsevier MIMET 00338 Review A r t i c l e

The control and measurement of fermentations

' C O 2'


Neil M. D i x o n a n d D o u g l a s B. Kell
Department of Biological Sciences, University College of Wales, Aberystwyth, Dyfed, UK

(Received 6 June 1989; revision received 18 August 1989; accepted 28 August 1989)

1. Introduction
It has been known for many years that the growth and metabolism of microorganisms is accompanied by the uptake and/or evolution of CO 2. To s t u d y and indeed to exploit the effects o f CO2 on microbial metabolism, it is necessary to control the level of (dissolved) CO2 within the culture medium. Whilst several articles cover one or two methods by which pCO2 may be measured and, hence, controlled, we know of no review of these. Thus, the purpose of the present article is to provide an overview o f the role, measurement and control of the magnitude o f p C O 2 during laboratory and industrial fermentations. We begin by describing the various ' C O 2' equilibria and the question of CO2-absorption rates.

2. 'C02' Concentrations
Although the partial pressure of C O 2 in the gas phase may be held constant, the ratios of the different possible species o f 'CO2' in the aqueous phase will vary as a function o f the p H and other factors. Since CO2 can hydrate and dissociate in water, the reaction scheme may be written [1] as:
CO 2 +

H 2 0 ~ H2CO 3 .-" HCO~ + H + '-" CO 2- + 2H +


In addition, it has recently been proposed that small concentrations of dimeric hydrogen carbonate ions (H3C20~-) exist near neutral p H [2]. Since the concentration of this species is negligible, however, such ions will not be considered in the following. At p H values o f < 8, the concentration of carbonate ions may be neglected [3] and only the following hydration reactions need to be considered:
CO 2 +

H 2 0 ~ H2CO 3 ~ HCO~ + H + •


Correspondence to: D.B. Kell, Department of Biological Sciences, University Collge of Wales, Aberystwyth, Dyfed SY23 3DA, UK. 0167-7012/89/$ 3.50 © 1989 Elsevier Science Publishers B.V. (Biomedical Division)

2.1. D i s s o l v e d C O 2 c o n c e n t r a t i o n

The concentration of CO2 in solution ([CO2]aq) is normally expressed by Henry's law [41:
[CO2]aq = K H P C O 2 (3)

where KH= Henry's law constant (mol.atm -1) and p C O 2 = the partial pressure of CO2 in the gas phase (atm). For cultures grown under atmospheric pressure, the proportionality of solubility and partial pressures (Henry's law) may be assumed without introducing appreciable errors [5]. At a temperature of 37 °C, KH=10 -1"61 [4] where the [COz] is expressed in molar terms. Thus, to obtain [CO2] in millimolar terms, K H = 10139. Hence:
[CO2]aq = 101"39 x p C O 2 = 24.6 x p C O 2.

In other words, when p C O 2 = l atm, the concentration of dissolved CO2=24.6 mmol. 1 -l.
2.2. B i c a r b o n a t e c o n c e n t r a t i o n in a p H - c o n t r o l l e d c u l t u r e w i t h a c o n s t a n t t e m p e r a t u r e a n d gas p h a s e

The equilibrium between CO2 and HCO~ is defined by a 'hybrid' equilibrium constant K 1' [4] where: K 1' I0 -pH" [HCO~]



From Eqn. 4, it follows that: log[HCO]] = p H - p K 1 ' + log[COz] • (5)

p K 1 ' is related to the thermodynamic p K of the reaction p K 1° and the ionic strength

I by:
pK~ ' = P K 1 ° - 0.5f(/)- bI.


From Davies's Eqn. 4:
f(1) = [11/2/(1 + 11/2) - 0.211 [298/(T+ 273)] 2/3


where T is the temperature in °C. I is the ionic strength of the medium and is given by [4, 6]:
l = V2 ECiZi2 (8)

where ci = the concentration of ion i and zi = the charge on ion i. To obtain the apparent PKa, l for the COz/HCO~ equilibrium, we use [4]:

max = KLaCs = KLaHpg. 5: log[HCO~] = pH -pKa. It is. c= concentration of the gas in the bulk liquid.157 pKa.f(I) - bI (9) Therefore. 1 = pKa. At 37 °C. time. and so from Eqn. From Eqn. we obtain PKa. (11) The absorption rate can be expressed in terms of partia pressures>fin gas and liquid (pg and Pl. 9. Hence: KLa= Rs/IO 1. 11: Rs = KL aH(pg . the rate of gas absorption (Rs) • vo1-1 of liquid is given [7] by: dc kLa (cs_C) Rs=d-t = h (10) where c s = gas concentration at the interface.1. It may be noted that these equations assume that there is an equilibrium between the p C O 2 in the gas phase and that in solution. HPl--0. kL = a constant dependent upon the diffusion coefficient o f the gas and h = the thickness of the stationary film. Then: Rs.Pl). it is usual to combine it with kL and write [6. 1+ log[CO2].39. because the value of h is not known. . 3. hence. Clearly. this implies that the rate of exchange o f CO2 between these two phases is rapid. 13.39when ['CO2'] is expressed in mM. a = interfacial area. it can be seen that: KLa = Rs/Hpg. In general. 7]: R s =Kea(Cs -c). (13) The rate of absorption of CO 2 is given by the slope of plots of 'CO z' concentration vs. (12) The maximum absorption rate for a given partial pressure occurs when c = 0 and. 1 . pertinent to determine the extent to which this is so. Rates of Exchange of CO2 between Fermentor Broths and Gas Phase Following a step increase in the p C O 2 in the gas phase. respectively) by making the substitutions c s = Hpg and c= Hpl (where H = Henry's law constant) into Eqn. from Eqn. H = 10 -1'61 when ['CO2'] is expressed in M [4] and H = 101.

P r o d u c t i o n o f C O 2 from various carbohydrate substrata is a useful diagnostic tool in the i d e n t i f i c a t i o n o f bacteria. Traditional Methods for Estimation of CO2 T h e detection o f CO2 in cultures includes b o t h qualitative a n d q u a n t i t a t i v e methods. 4. 9] a n d the a p p e a r a n c e of cracks in agar m e d i u m [10]. d e s o r p t i o n a n d solubility of CO2. see also H o et al. 71 75 103 . in addit i o n to some o f the reactions involving CO2. D u r h a m tubes inserted into b r o t h cultures [8]. The Eldredge tube [11]. e.g. we will n o w discuss various m e t h o d s by which CO2 m a y be estimated (Table 1). 22 12..g.158 For a detailed review of the diffusion. barium carbonate) Decrease in volume of a gas sample after chemical removal of CO2 Increase headspace volume above an acidified sample Increase in pressure of headspace above an acidified sample Differences in thermal conductivities of gases Differences in mass to charge ratios of ions Absorption of IR radiation Changes in light transmittance of pH indicator Differences in column retention times of gases Changes in pH induced by CO2 Changes in water conductivity caused by CO2 Flow of electric current due to reduction of CO2 Change in resonant frequency due to adsorption of CO2 Selected references 20 21. T h e rate o f diffusion o f CO2 t h r o u g h a silicone m e m b r a n e m a y also be used to measure pCO2 [7a]. Several m e t h o d s for detecting CO2 product i o n have b e e n used. d i s p l a c e m e n t o f agar plugs [8. METHODS ARE LISTED IN THE ORDER THAT THEY APPEAR IN TEXT Method Titrimetric Colourimetric Gravimetric Volumetric (Orsat) Volumetric (Warburg) Manometric Katharometer Mass spectroscopy IR Fibre optic probes Chromatography Potentiometric Conductiometric Amperometric Piezoelectric Principle A standard alkaline solution is neutralised after absorbing CO2 Modification of colour of an indicator solution Isolation and weighing of a compound of CO2 (e. H a v i n g o u t l i n e d C O 2 . which is a device for t r a p p i n g CO2 in b a r i u m hydroxide. 20 27 31 31 27 35 27 48 36 50 70. [6]. can be used for b o t h the qualitative (observation o f a precipitate) a n d q u a n t i t a t i v e (recovery of the TABLE 1 METHODS USED IN DETERMINATION OF CO2..a b s o r p t i o n rates a n d equilibria.

22]. do not require titration of the CO2-absorbing solution.3) at which point the excess of barium hydroxide is neutralised [16. In the estimation of CO2 by the Conway microdiffusion method [16. Similarly. The CO 2 in the gas sample is absorbed into a standard barium hydroxide solution which. in the form of a square of flat ground glass.1. Colourimetric methods. Displacement of agar plugs is used in the hot-tube method which is more sensitive than the hot-loop method [15]. 17]. The cell consists of a Petri dish in the centre of which is fused a piece of glass tubing to form an inner cell or compartment.3 [19]. The insertion of a red-hot loop into cultures resulted in a copious evolution of gas [13. is titrated with standard hydrochloric acid [20]. There have been a number of designs of the apparatus for the microdiffusion method [16-18] but they are all based upon the standard cell illustrated in Fig. for the estimation of the CO 2 concentration in a gas. However. 14] and this method of CO2 detection is known as the hot-loop method. CO 2. Free CO2 reacts with the sodium hydroxide to form sodium bicarbonate. liberated from the sample in the outer compartment. may enclose completely any gases within the dish. Titrimetric methods may also be used for the determination of CO2 in gases. . In the hot-tube method. A mixture o f vaseline and paraffin is smeared onto the lid to assist the seal. The contents of the inner compartment are then titrated with a standard hydrochloric acid until the thymolphthalein indicator is just colourless (pH ~-. CO2 in a liquid sample may be estimated directly. The edge of the outer dish is ground so that a lid. The colour intensity of a solution of phenolphthalein's sodium salt decreases as the concentration of CO2 increases [21. Conway's standard microdiffusion cell. 1.159 precipitate) detection of C O 2 [12]. these procedures depend upon the evolution of gaseous CO 2 from the growth medium which does not always occur due to the high solubility of CO2 in water. an agar plug is displaced by the evolution of gaseous CO2 when the culture is heated to 80°C. after the addition of phenolphthalein indicator. 9. CO 2 is liberated by acidifying the sample. the concentration of CO 2 may be estimated by the modification of the colour of alizarin yellow R in the presence of NaOH [22]. is allowed to diffuse to the central compartment where it is absorbed by a standard barium hydroxide solution [KOH or NaOH may be used instead of Ba(OH)2] containing thymolphthalein indicator. 17]. CO2-detection tubes are also based on a colour change caused by a chemical reaction between the gas and tube / DISH CPE Fig. Completion of the reaction is indicated by the development of the pink colour characteristic of phenolphthalein at the equivalence point of pH 8. Phenolphthalein is added to the sample which is then titrated with a standard sodium hydroxide solution.

Enzymatic methods for the estimation of various compounds are specific for the compound(s) involved. 2 8 . The pH may then be determined by the use of indicators [25] or with a pH electrode (see also Section 9. 2.1. 2. however. the pH is a measure of the concentration of CO 2 [24]. interference due to organic or inorganic compounds commonly found in fermentation liquors are either negligible or easily eliminated [31]. When a gas containing C O 2 is brought into equilibrium with sodium bicarbonate solution. Other traditional methods of measuring CO2 involve the passage of a measured stream of gas through absorbants followed by gravimetric [e. 27]. When combined with a spectrophotometer the pH.). following the removal of CO 2 by its absorption into KOH. 27]. Designof an apparatus for manometricdeterminationof C O 2. of a sodium bicarbonate solution. may be estimated on a continuous basis [25]. at constant pressure.g. C O 2 is estimatedby measuring the increase in pressure of the headspace followingacidification of sample [31]. was measured and corresponded to the p C O 2 [26.1.32]. Although such a method requires only very simple instrumentation and may easily be automated. it is not able to provide genuinely continuous monitoring of the total 'CO2'. Total 'CO2' may be measured by first saturating a sample with CO2.. Malate dehydrogenase then ACIDINLETTUBE GASOUTLET TUBE SAMPLE I TUBE o o oo MANOMETER ° o J C~ T~ M S A r D~ A I T I R O T TE SAMPLE I DISCHARGETUBE ~ ~ _ PUMP '~ --~CO2 INLETTUBE Fig. and hence p C O 2. Interference from volatile acids. . and an example of an apparatus for such a method is shown in Fig. One such method exists for the estimation of 'CO2' and is commercially availabe (e. is a possibility. containing indicator and sparged with a CO2-containing gas. Such a method was employed in the Orsat apparatus in which the change in volume. from the Sigma Chemical Company).. 12] or volumetric measurements [26. of a gaseous sample. which become gaseous in acidic conditions.g. then acidifying the sample and measuring either the change in the volume of the headspace at a constant pressure or the change in the pressure of the headspace at a constant volume [26. The method is based upon the reaction catalysed by phosphoenolpyruvate carboxylase which produces oxaloacetate from phosphoenolpyruvate and 'CO2' [33].160 contents [23]. The latter method does not require gas meters for the accurate measurement of volume.

o f the two gases.g. since they are performed on samples taken from a patient [e. Sample and reference cells each contain a Pt wire ( A . then Wires B and D will lose more heat than will Wires A and C. Under these conditions. the bridge is balanced and no current flows through the galvanometer. each wire will attain the same temperature and resistance.161 converts N A D H and the so-formed oxaloacetate to malate and NAD +. The behaviour o f the katharometer reflects the thermal conductivity o f the gas (mixture) which is passing through it. Two of the cells (B. each containing a glass-coated Pt wire identical with the others.. 3. CO2 Analysis Katharometer by The principle of C O 2 analysis by katharometer [27] is as described below. C) are exposed to the sample gas. 34]. When the bridge current is constant and all four cells are exposed to the same gas. To measure the p C O 2 of the effluent gas from a fermentor the constituent gases ---ql A I) C I 44A. the 'traditional' methods. are introduced into the cells (air into the reference ceils and the mixture into the sample cells).g. forming the arms o f a Wheatstone bridge. and in this instance pCO2. 32. The consequent change in the conductivities of the wires will unbalance the bridge and cause a deflection of the galvanometer. A simplified representation o f a katharometer. each of which is identical. can only be used for spot checks and are therefore not suitable for continuous analysis. These four wires form the arms of a Wheatstone bridge (Fig.' B~__ Fig. The principle of CO 2 measurement using a katharometer is as described in text. However. at either 340 or 380 nm. discussed above. 3). . blood) are also unsuitable for continuous use. With the exception o f the colourimetric method. Within the katharometer are four small cells. When the absorbance is read. for CO 2. 5. Methods employed in the gas analysis of clinical samples (e. e.g. D) are exposed to a reference gas and the other two (A. If two gases of different thermal conductivities. the decrease in absorbance upon the addition o f P E P is proportional to the original 'COil content of the sample [33].. such as air and an air/CO2 mixture.D ) . the measurement of gases by katharometers is not specific. with the size of the deflection being dependent upon the difference in the thermal conductivity.. This is because the thermal conductivity of air is greater than that of CO2.

according to m/z = (H2r2)/2 V where H is the strength of the magnetic field.40]. the mass spectrometer can be adapted for selected ion monitoring to permit highly selective determinations of very high sensitivity [36]. These nonmagnetic mass spectrometers employ as a mass filter four electrically conducting parallel rods precisely located in a ceramic holder. made of a gas- . 180 ° out of phase with each other [35. At a given RF/DC voltage. for example. seriously compromising the quantitation of the individual species [37]. pass through the filter.5). 38-40]. Despite this problem.41]. By continuous scanning of either the magnetic field or the accelerating voltage. Ions entering at one end of this array are constrained to make oscillations in the transverse electric field. The reason for this is that mass spectra tend to be complicated and examination of several compounds simultaneously can lead to significant overlap. mass spectrometry cannot always be used to examine a number of substances simultaneously without additional separation techniques such as gas chromatography (see Section 1. and the ions. The opposite rods are connected electrically and a radio-frequency (RF) voltage superimposed on a DC voltage is applied to the two pairs of rods. are accelerated by an electric field through a series of slits to form an ion beam. is covered with a plastic membrane and immersed into the fermentation medium. r the curvature of the pathway (instrumentally dictated) and Vthe voltage used to accelerate the ions [35]. Since the diffusion coefficient is dependent upon both the gas and the membrane used. This ion beam is then deflected by a magnetic field. This method involves passing a carrier gas through a tube. 38-40]. by exploiting the fact that different gases have different isotopic compositions.. by gas chromatography as described in Section 8. e.162 would first have to be separated. In addition to the identification of compounds. the membrane inlet is selective. only ions of a specific m/z value will avoid collision with one of the rods and will. ions of different m/z values can be focussed on the detector and subsequently measured. in fact. Another means by which the dissolved gases in fermentation media could be sampled is by a carrier gas tubing method [3. 42]. 6. CO 2 is measured at m/z=44. The sample is ionised. Mass Spectrometry The basic principle of mass spectrometry is the separation and registration of ionic masses [35]. H. The only possible interference is from N20 (formed by many denitrifiers) and from volatile fatty acids (which may break down to CO2 within the mass spectrometer [39]). with a mass-to-charge ratio of m/z. Scanning the RF and DC voltages may be performed rapidly with the ions being transmitted sequentially in order of their m/z ratios with constant resolution [35. gases produced by fermentation processes can. thus. by electron bombardment. Sampling dissolved gases is achieved by a membrane inlet to the mass spectrometer [ 3 9 . be estimated simultaneously by mass spectrometry [38. or across a membrane. or porous material. A vacuum is applied to the inside of the probe and the dissolved gases diffuse through the membrane. A perforated capillary or sheet. However.g. The type of mass spectrometer described above is the magnetic mass spectrometer which is perhaps more often used for the analysis of fermentor exhaust gases whilst quadrupole mass spectrometers are used for the estimation of dissolved gases [39].

and a second fibre carries the transmitted signal (Fig. However. proportional to the absolute pressure of the gas in an analysis cell [27]. acetaldehyde and substances with . The development of fibre optic pH probes [46] has led to the production of fibre optic CO2 probes [47. ambient pCO2 controls the pH of a bicarbonate buffer solution. When analysing a single component in a gas stream. 48]. The shutter position is a measure of the pCO2. therefore. a number of new uses of fibre optics in sensing have been proposed.S H bonds [44]. Since this absorption is characteristic of . interfere. Unfortunately. methods such as mass spectrometry and infrared analyses (see the following section) are normally prohibitively expensive for routine use on single fermentors at the laboratory scale [43]. 51]. 7. not selective for CO 2 unless the gas sample has been treated to remove interfering substances prior to the pC02 measurement and. 44]. therefore. thus. IR gas analysis is. 4) to a receiver which converts the signal to an electrical since organic acids or bases that are able to penetrate the membrane may also cause such a change in pH [50. this difference is used to actuate a servo system.163 porous material. The optical approach avoids problems of making miniature glass or liquid membrane pCO2 electrodes [49] with adequate reference electrode stability. Water absorbs strongly in the IR [45] and. Optical Methods of CO 2 Determination 7. In a null-balance instrument. The principle of IR spectroscopy for estimating the partial pressures of gases is as follows [27. after passing through the absorbance cell through which the gas is flowing. The difference between the two signals is a measure of the pCO 2. are not necessarily due solely to a change in PfO2. as stated above. a so-called 'nondispersive analyser' is used. which absorb in the IR and those may. The intensity of the transmitted light is compared with the intensity of the incident light by using a 'blank' or reference cell filled with CO2-free gas.L IR absorbance The absorbance of IR radiation by a gas (mixture) is proportional to the mass of absorbing molecules in the light path and is. Fibre optic measurements Stimulated by recent developments in fibre optics technology. immersed in the fermentation medium. . contained within a membrane.2. the changes in pH.15/zm. the presence of water vapour in the gas sample may interfere with the IR spectrophotometric measurement of CO 2. This examines absorbance in the wavelength range of 3 . Some other compounds from bacterial cultures.OH bonds. 7. are acetone. hence. which then influences the optical transmittance of a colourimetric pH indicator. In such fibre optic CO 2 probes. This adjusts a shutter situated in the light path of the reference cell until the sample and reference radiations are matched. this technique may be rather expensive. passes to a detector that is selective for the wavelength at which the component being analysed absorbs. volatile alcohols may also interfere. One fibre carries light to the buffer/indicator solution. that are measured with such a device. The exhaust gas from the tubing is then analysed. The light beam.

g.164 TEFLONSHEATH INDICATOR IN SILICONE TUBE ~-~ ... 54. the retention time. The retention times of the components are then calculated relative to that of the internal standard. After introduction onto the column. they are only of use on a noncontinuous basis. C h r o m a t o g r a p h i c C O 2 D e t e r m i n a t i o n In gas liquid chromatography. TO DETECTOR ~. 4. Cramers and McNair [54] and Willett [55]. Construction of sensor tip of a fibre optic probe. Perry [52]. Various components of a sample can be separated only if they have different K values.g. a sample is introduced into a stream of carrier gas (the mobile phase) and is swept through a chromatographic column which contains a stationary phase. KHCO3 and KC1 [48]. Provided that the detector and separation procedure are appropriate. is defined [36.. However. thermal conductivity detector.. temperature and carrier-gas flow rate. each peak will correspond to one of the components of the original sample.. The time taken to pass through a column. the lower the solubility of the solute in the stationary phase. Grob [531. The area beneath each peak can then be correlated to the concentration of the corresponding component in the injected sample [36. A pieceof silicone tubing is then plugged at one end with white silicone adhesive and filled with a solution of phenol red. see. .. . Larsson and O d h a m [36]. will be affected by the type of column. Two fibres extending from end of appropriately opaque and inert tubing are sealed into it with epoxy. i. 8. .. mass spectrometer) connected to a recorder or a printer/plotter and are consecutively registered as chromatographic peaks. . The lower the value of K. e.. the components enter a detector (e. EPOXY " SILICONE ADItESIVE OPTICHBRE Fig.. . Clear siliconeadhesiveis coated overepoxyexceptat fibre ends.. Once separated. the quicker it will pass through the column. For a more detailed discussion of chromatographic separation theory. 52. since chromatographic techniques rely upon the spatial separation of the compounds that are being quantified. the sample components are allowed to distribute between the stationary and mobile phases according to their partition coefficients. K. An internal standard may be added to the sample before analysis to compensate for possible variations in analytical conditions. .e.Epoxy is also layered onto cut end of fibres to prevent binding of indicator dye. 52] as: K=Cs/Cm where C s is the concentration of solute in the stationary phase and Cm is the concentration of solute in the mobile phase. 55]. the partition coefficient. . .

. However. 5. the use of the Severinghaus-type potentiometric electrode for measuring the levels of dissolved CO2 is hampered by its slow response time. Pt(s) and: K2CO3(s). complex relationships among multiple chemical species during unsteady diffusion were shown to cause hysteresis and pH-dependent response rates [59]. 5). Such a steam-sterilizable Severinghaus-type p C O 2 electrode that can be calibrated in situ.CO2 is absorbed into or desorbs from buffer solution and corresponding change in pH is measured by pH electrode [50]. therefore.g. 02 and CO 2 [56-58].aq by ion-selective potentiometric electrodes. has. 'invasive'. including gases. the changes in pH that are measured with such a device are not necessarily due solely to a change inpCO2. Digital simulations may be used to predict the dynamic response of such electrodes [60] and electrode stability may be sacrificed to improve response time. 59].. been described [61]. additionally. The fermentation medium and the buffer solution within the electrode are separated by a membrane which is 'freely' permeable to CO2. As with the fibre optic probes mentioned above.C a O ( s ) can be used for CO2 determination in air. such as the Severinghaus-type electrode. involves the measurement of the pH of a thin layer of a buffer solution in equilibrium with ambient CO2 [57.165 9. such as the Severinghaus-type (Fig. C O 2 E l e c t r o d e s 9. CO2(g). It was shown that the triple contacts: K2COs(s). Potentiometric Ion-selective electrodes may be utilised for the estimation o f a wide range of compounds. is that they must be immersed in the fermentation broth and [50]. CAP pH ELECTRODE - ~CE BUFFERSOLUTI01~ ~RODE SILICONE RUBBER MEMBRANE ~ING Fig. The determination of pCO2. Design of Severinghaus-typeelectrode. ammonia. The use of such invasive devices leads to a number of potential problems: an additional port on the fermentation vessel will be necessary leading to an increased risk of contamination: the device must be robust enough to withstand sterilisation: the device is required to be calibrated easily without having to be removed from the fermentation vessel. in fact.or in O2-bearing gases at temperatures ranging from 450 to 750 °C [62]. e. K2CO 3 may be used as a solid electrolyte for the potentiometric measurement of gaseous carbon oxides.1. Z r O 2 . the electrodes are by no means cheap. Another problem with devices. CO2(g).

66 but cf. therefore. 71].. The principle involved is once again the measurement of the pH change within a buffer solution and. are invasive devices and. caused by the absorption of CO2 when the effluent gas was bubbled through a solution of sodium carbonate. [3. Amperometric In contrast to CO2. The back wall of the thin water layer (see Fig.65]. 0 2 is routinely and more-or-less reliably measured either . possess all of the problems associated with invasive devices listed above.1. A potentiometric method was proposed whereby changes in pH.166 The fabrication and performance of a variety of ion-sensitive field-effect transistors (ISFET). 71]. Electrodes positioned in the water layer continuously measured its conductance. 9. However. for gas analysis. Noninvasive methods for determining pCO2 have received attention for clinical uses in the transcutaneous analysis of arterial p C O 2 [e. however.3. however. 6) was a porous screen separating the water from a mixed-bed ion-exchange column which continuously removed ionic species from the water layer. rather than the determination of the concentrations of particular ions in a solution. The diffusion of CO 2 into the water layer and the removal of ionic species by the mixed-bed ionic exchanger established a steady-state concentration gradient of CO 2. ISFETs. like Severinghaus-type electrodes. This method is still hampered by the problems associated with Severinghaus-type electrodes listed above. The ISFET has attracted considerable interest because it is envisaged that a single miniaturised solid-state chip could contain multiple sensors and be used to sense several ions simultaneously. The response times. such a technique is of use in control systems for maintaining a normal arterial pCO2 by artificial respiration since such sensors are capable of continuous measurement. eliminating the need for the more technically demanding measurement of pCO2. 9. A continuous conductimetric sensor for CO2 (Fig. However. The gradient was proportional to pCO2 in the gas phase and the cell conductance was proportional to (pCO2) 1/2 [70. were much slower than with Severinghaus-type electrodes and the ISFETs still responded to organic acids.1. which are in effect solid-state devices (no internal reference electrolyte). The measurement of pCO2 in the effluent gas is said to give an excellent approximation of the pCO2. have been described [ 6 3 .01 atm. despite the slow response time. 6]. 67-69].aq and allowing the utilisation of noninvasive devices. ISFETs could be used as noninvasive devices by simply placing them in the effluent gas flow. When field-effect transistors were coated with PVC or silicone rubber membranes.g. larger responses to CO2 than with ion-selective membranes were observed and the resulting ISFETs behaved very much like Severinghaus-type potentiometric CO2 sensors [51]. Conductimetric Conductimetry has also been applied to the measurement o f p C O 2. were measured [23]. It has been shown that ISFETs based on ion-selective membranes were subject to positive interference by either CO2 or organic acids [51]. 9. Invasive and noninvasive electrode methods for determination of CO 2. this device was evaluated for use in the determination of pC02 only up to 0. 6) was described in which CO 2 diffused through a hydrophobic gas-porous membrane into a thin layer of pure water [70. to use an invasive device in a noninvasive way. It is possible.2.

GOLD CONDUCTIVITY ELECTRODES Fig. paramagnetically [27] or by means of the Clark-type amperometric electrode [43. 6. . When an appropriate potential is applied across the electrodes. Such electrodes are usually constructed in Perspex. Design o f the Clark-type electrode. 72] which consists typically of an Ag ring anode surrounding a Pt cathode (Fig.B E D ION H--I L EXCHANGER GLASS TUBING _ _ WATER-POROUS SPACER SCREEN WATER-POROUS SUPPORTING SCREEN .M I X E D . 02 is reduced at the cathode in the reaction: 2 H + + 2 e . 7.+ 1A0 2 -- H20. An Ag ring anode surrounds a Pt cathode and the circuit is completed by a KC1 solution. Schematic representation of a continuous conductimetric C O 2 sensor.- 167 THIN WATER GORE-TEX GAS-POROUS . Diffusion of CO 2 through the membrane into water layer and removal of ionic species through screen by ion exchanger establish a steady-state gradient of CO 2 [70. 3M KC1 _ TEFLON MEMBRANE Pt CATHODE -- "4- Fig. 71]. 7). Electrodes positioned in thin water layer measure its conductance.MEMBRANE PLATINISED.

CO2 may also be reduced at such a cathode in the following way [74]: in anhydrous conditions.HCO5 + O H .C O + C O 2-. possible [75]. an amperometric CO 2 probe is. As potential difference between Pt cathode and Ag anode is increased. 7).HCO~.CO5. 8) [73]. H evolution commences and there is a large increase in current. The amount of current is.CO~. 2 C 0 2. . Thus. As the potential of the cathode is made more negative. if 0 2 is present. This electrode for amperometric CO2 measurement was a three-electrode system (Fig.168 The circuit is completed by a KCl-salt bridge between the anode and the cathode and the anode reaction is: Ag + C1 . a current flows and AgC1 is deposited on the anode. lower.-. Hence. in the presence of water. Current-voltage curve for reduction of 0 2 in Clark-type amperometric electrode (Fig. If potential difference is increased above plateau region. thus. consisting of AIR-SATURATED / -1~0 iV) 1/2 AIR-SATURATED WATER 0 -0-5 POTENTtAL DIFFERENCE Fig.AgCI + e . .-. therefore. therefore. C O 2 q. producing a plateau region (diffusion-limited current) (Fig..-. rate of O 2 reduction at cathode increases. CO5 + H20 -.e . If less O2 is present its diffusion is slower and the plateau current is. OH. 8. C O 2-b e .-. O2 is reduced more rapidly until a potential (difference) is reached where the reduction rate is equal to the rate at which O2 can diffuse to the cathode. 9). I f the potential difference is increased beyond the plateau region the evolution of H commences at the cathode and there is a large increase in current. A plateau is reached when reduction rate of 0 2 is equal to rate at which O z can diffuse to cathode. linear with the O2 concentration at the electrode surface.+ CO5 -.

g. A number of processes has been reported for the reduction of CO z although these have been directed primarily to aqueous solutions [e. current passing through the reference electrode may cause its potential to deviate and give rise to internal polarisation. thus. 7). Interference by the chloride ion often arises when mercury electrodes are used and may be avoided by using nitrate or perchlorate salts.78]. Since Perspex dissolves in DMSO. a working electrode. Design of an amperometricCO 2 electrode[102]based upon the design of an O2/CO 2 electrodeby Albery and Barron [75]. a counter electrode and a reference electrode.. the electrodewas constructed from nylon. However. 02 is much more readily reduced than is CO 2 and. This. and resistive drops due to the presence of junction potentials across solutions and the interface between different materials.13 V vs. Potassium nitrate is an alternative when chloride cannot be used. at the potential required for the reduction of CO 2 (at an Ag working electrode the ½-wave potential is -2. 79]. This can be avoided by the use o f tetraalkylammonium salts. leads to a decrease in the absolute potential difference between the working electrode and the reference electrode. inert supporting electrolytes are used [76.169 TEFLONMEMBRANE I / DMSO+ ELECTROLYTE "0"-RING P~I~RENCE ELECTRODE NORK/NG ELECTRODE COUNTER ELECTRODE ELECTRODES REFERENCE COUNTER ELECTRODE ELECTRODE Fig. Addition of the third electrode mimimises these problems as the current now flows between the working electrode and the counter electrode with the potential of the working electrode being set relative to that of the reference electrode [76. However. 9. if present. as compared to the Clark-type electrode which is generally a two-electrode system (see Fig. the foot of its polarographic reduction wave may extend to values sufficiently positive as to cause unwanted interference [79]. To reduce the resistive drops to an insignificant level. In such twoelectrode systems. the standard calomel electrode [75])water is electrolysed [84]. will give rise to the production o f superoxide which accumulates and interferes with the measurement . Potassium chloride is commonly used as a supporting electrolyte since it is easily available in high purity form and the mobility of the potassium ion and the chloride ion are almost exactly equal.83]. 8 0 . when a high concentration of potassium ions is present. contra-indicating the use of aqueous solutions within the electrode for analytical purposes.

1. Although the electrode produced an output with a linear relationship to 'CO2'. The principle of this is that 02 is reduced. 10. 10. at the metallised membrane in an aqueous solvent to produce water. The 'CO2' concentration was estimated from the decrease in the output current from the 02 electrode. the presence of CO2 results in a reduction. Reduction o f CO 2 mediated by an electrocatalyst.3. proportional to the pCO2. this problem was circumvented by the use of a metallised membrane as an electrochemical filter for the removal of 02. a great deal of attention has been devoted to the utilisation of CO2 as a source of C in the elec- ANODE .1 0 min. CO 2 permeates through the dialysis membrane and is assimilated by the bacteria. In the electrode described by Albery and Barron [75]. Determination o f p C O 2 with an amperometric electrode should be able to give continuous measurements. the temperature range was limited and the response time was 5 . Such electrodes generally exhibit a rather poor dynamic range. especially. However. Cross-section of a CO 2 sensor which utilises immobilised thermophilic bacteria [85]. in the case of an electrode intended for the monitoring of pCO2 in anaerobic fermentations the presence of Oz will not be a problem. caused by the decrease in the 02 concentration corresponding to the increased bacterial respiration. As with the immobilised bacterial sensor.. in doing so. The construction of this CO2 sensor is as shown in Fig. pH and pCO2 [50]) and be both simple and cheap to construct.E~C~OL~E DIALYSIS MEMBRANE CATHODE BUFFER -* BUFFER GAS ] MEMBRANE IMMOBILISED BACTERIA 0-RING Fig. It is a decrease in 0 2 concentration that is measured by the electrode. Recently. a CO 2 sensor using immobilised thermophilic bacteria and an amperometric 02 electrode was described [85].170 o f p C O 2 [75]. their respiration rate increases resulting in a decreased 02 concentration. at a lower potential than that which is required to reduce CO2. The immobilised bacteria were supplied with O2-saturated buffer containing various metal ions and 200/~M glucose. The bacteria assimilate CO 2 and.. . causing their respiratory rate to increase [86]. CO2 may be detected by another method which utilises an 02 electrode [87]. of the current output of the 02 electrode. 9. the range of CO2 concentrations tested was small. Recently. if it is to be used under anaerobic conditions. have a linear relationship between output and pCO2 (as compared with the Severinghaus-type electrode which has a logarithmic relationship between the variable actually measured.

IR methods probably suffer too many interferences to be of general use in fermentations though the microwave region does not yet seem to have been explored. Such interactions would 'delay' CO2 desorption from the solvent giving rise to higher concentrations of 'CO2' than would normally be encountered and reflected in the increased current. developed for such CO2 utilisation. Fibre optic probes seem to offer many advantages whilst continuing technological development work may be expected for potentiometric devices in particular. ease of use and expense. or homogenous. .52 V [83. CO + H20. response time. there is likely to be a trade-off between resolution. Although the use of Re(bipy)(CO)3C1 in an electrode may in theory appear to be beneficial.g. The methods discussed herein. in practice. e. the use of other electrocatalysts in amperometric electrodes may improve this method of CO2 measurement.g.. polyelectronic reduction of CO2 may occur at much less negative potentials.. cover the spectrum of simplicity. On the other hand. and summarised in Table 2. However.2 V [74. Re(bipy)(CO)3C1 was reported to reduce CO2 to CO at a potential of -1. In a system based on Re(bipy)(CO)3C1.-. The purely photochemical systems may be classified as either heterogenous. When Re(bipy)(CO)3C1 was used as a catalyst in the electrochemical reduction of CO2 to CO [83]. the E01for the dielectronic reduction of CO2 to CO in aqueous solutions at pH 7 is only -0. 80. precision. That the use of the electrocatalyst in an electrode caused an increase in the time taken for the current to decrease following a decrease in the pCO 2. requiring potentials as negative as . 96] and occurs as follows: CO2 + 2H ÷ + 2e. worsened matters by reducing the rate of current decay after a reduction in pCO2 [102].. organic dyes or transition metal complexes [96-99].25 V. Direct noncatalysed reduction of C02 follows a monoelectronic pathway (equations given above). high yields of CO were photogenerated from CO 2 [100]. to reduce the potential required for the reduction of CO2. substantially below the potential for monoelectronic CO2 reduction. 101]. Concluding Remarks and Summary As with any set of analytical methods. One possibility would be to use an electrocatalyst. 88]. this electrocatalyst offered no added benefit and. mass spectrometry is probably the most widely used method industrially and/or where money is no object. employing aqueous solutions of metal ions [82. the system displayed high current efficiency and long-term stability (high overall turnover).171 trochemical synthesis of organic products [e.and photo-electro-chemical). was possibly due to interaction between CO 2 and the electrocatalyst. photo. making use of semiconductor suspensions [89-92]. in fact. few catalytic systems are known despite the development of a number of different strategies (electro-. 93-95]. 10. in an amperometric CO2 electrode. At present. However. Such catalytic properties would appear to make Re(bipy)(CO)3Cl an attractive catalyst for use in an amperometric CO2 electrode for the improvement of selectivity and specificity. accuracy.

H. W.g.s e n s i t i v e p i e z o e l e c t r i c crystal which gave a linear response to pCOz in the range of 0. pp.F i l h o et al.01 atm Simple and cheap. Berlin. 19. [103] h a v e r e c e n t l y d e s c r i b e d a C O z . A. UK.K. Biotechnol. (1985) Potentiometric titrations of aqueous carbonate solutions.16 atm under fermentation conditions. Bioeng. References 1 Knoche.018-0. 3-11. e. 2 Covington. C. organic acids/bases may interfere. E (1977) Desorption of CO/from fermentation broth. Springer-Verlag. Noted added in proof F a t i b e l l o .. Simple and cheap. . Interferences not known but volatile interferences likely Mass spectrometry IR Fibre optic probes Gas chromatography Potentiometric Dissolved gases or fermentor exhaust gases Exhaust gases Dissolved 'CO 2' Overcomes problems of making miniature glass electrodes Exhaust gas or small liquid samples Dissolved 'CO 2' Conductiometric Amperometric Piezoelectric Exhaust gases Exhaust gases Exhaust gases Acknowledgements We are grateful to the Biotechnology Directorate of the Science and Engineering Research Council. Cumbersome Specific for 'CO 2' but unsuitable for continuous use Not specific for CO acids or bases may interfere Unsuitable for continuous use Slow response time. In: Biophysics and Physiology of Carbon Dioxide (Bauer. 265-281. eds. logarithmic response Evaluated only up to 0. Soc.. G. 14. for financial support. GC Expensive but precise and accurate Expensive. 801-819. (1980) Chemical reactions of CO 2 in water.. Rev. Gros. and Barrels. Should ideally be combined with a separation method.172 TABLE 2 SUMMARY OF METHODS FOR ESTIMATION OF 'CO z' DESCRIBED IN TEXT Method Titrimetric Colourimetric Gravimetric Volumetric Enzymatic Katharometer Applications Liquid samples Liquid or gaseous samples Exhaust gases Dissolved 'CO 2' Dissolved 'CO 2' Detection of gases Comments Unsuitable for continuous use Unsuitable for continuous use Unsuitable for continuous use Not able to provide genuinely continuous monitoring. 3 Yagi. electrodes not cheap. and Yoshida.). H. Not specific for C O 2 Measures pH . Chem.

W. Exp. 81. Princeton. Bakteriol. J. Eng.. Chem.N... R. A.D. London. H.. (1959)Colorimetric Methods of Analysis. Infektionskr. E D.L. Dairy Res. O. Agric. K.-D. T. 22 Snell. H. I. W. L. 27 Elsworth. J. Edward Arnold. Crosby Lockwood & Son.A. Y. R. 170-224. and Beauman.D. and Neill. (1982) CO 2 Equilibria and Their Applications. 59. Eng..E.). Methods. (1914) The bacteriology of cheese of the Emmenthal type. Mitchell. S. New Jersey. 12 Williams. Gen. 681-699. Wiseman. Microbiol.A. J. W.T. 28 Van Slyke.. In: Methods of Quantitative Micro-Analysis (Milton. 134-136. APHA-AWWA-WPCF. Iowa Agric. 7 Pirt. (1945) The formation of carbon dioxide by lactic acid bacteria and Bacillus licheniformis and a cultural method of detecting the process. 18 Obrink. J. eds. Cambridge University Press.. H. and its application to an in-line carbonation meter.E. J. T. R.. spec. London. W. M. R. Addison Wesley. Washington. 229-420. J.H. Cambridge. J. J.. Van Nostrand Co. R. London.M. D. 24 Snell. 21 Milton. Food Technol. G. (1970) The measurement of oxygen absorption and carbon dioxide evolution in stirred deep cultures. B. (1955) Gasometric methods of micro-analysis. Burgess Publishing. M. In: Methods in Microbiology 2 (Norris.D. Res. and Franson. (1961) Colorimetric Methods of Analysis. (1976) Hot-loop test for the determination of carbon dioxide production from glucose by lactic acid bacteria. and van Vuuren. (1982) Gas solubilities in microbial culture media. D.J. 8 Hayward. 5. and Abdel-Malek. London. E (1923) Studies on Streptococcus paracitrovorus group. J. 6 Ho. Blackwell Scientific Publications. J. 273 . 14 Sperber. 14. Oxford. 990-991. 35. Hyg. 17 Conway. eds. M. E. Edward Arnold. 541-545. 35-44. Appl. 16 Milton. London.173 4 Butler. B. L.E and Waters.. (1955) A modified Conway unit for microdiffusion analysis. Bacteriol. and Snell. Res. 23 Rozzi. 13 Hammer. Third Edition. J. 11 Eldredgde. J. A. pp. Academic Press. Eng.N. (1983) Potentiometric method for the determination of carbon dioxide in biogas. M. D. (1987) A modification of the hot-tube method for the detection of carbon dioxide produced by heterofermentative Lactobacillus strains. 5 Schumpe. Smith. 26 Wilson. 19 Greenberg.. E and Waters. J.275. (1957) Detection of gas production from glucose by heterofermentative lactic acid bacteria. R. Connors.W. (1924) The determination of gases in blood and other solutions by vacuum extraction and manometric measurement. M. J. Van Nostrand Co. 16. Adv. Stn. D.E and Water. Minneapolis.A. J. and Johnson. 306. W. 1-38. J. Quicker. and Deckwer. D. and Ribbons. (1953) Species differentiation of oral lactobacilli from man including descriptions of Lactobacillus cellobiosus nov. Biotechnol. Abt. 29 Dixon. Microbiol... A. and Swan.). London. E. K. . and Rogers. M. 505-512. 28. 2. 9 Gibson. W.. 30 Umbreit. 6. Edward Arnold. (1959) A Text-Book of Quantitative Inorganic Analysis Theory and Practice. and Campbell. Biochem.S. W. eds. A. Vol. Burton. Meet. Proc. MBAA San Diego. D. 40.W. Microbiol. C. R. Princeton. Home Econ. Zentralbl. and Hawkes. L. New Jersey. 521-605. Jenkins. 24.). Vol. In: Methods of Quantitative Micro-Analysis (Milton. C. 213 -228. Minnesota. K. A. pp. Environ. Biochem. and Shanahan. pp. Longmans. E (1957) Manometric Techniques. 523-573. 83-125. and Stauffer. and Baker. R.).A. 61. 9-15. (1952) Continuous photometric determination of carbon dioxide in gas streams. 15 Dicks. W.T. 65. C. (1952) Manometric Methods as Applied to the Measurement of Cell Respiration and Other Processes. Chem. A. (1952) The detection of heterofermentation by lactic acid bacteria. Bull.J. E. Burris. A. E (1955) Illustrative examples of microvolumetric procedures. 5-21. IIA. M. R. Disraely. (1962) Microdiffusion Analysis and Volumetric Error. 24. Biochem. Green & Co. 25 Maxon.W. E (1987) Carbon dioxide transfer in biochemical reactors. Parasitenkd.J. 10 Rogosa. London. (1981) Standard Methods for the Examination of Water and Wastewater. In: Methods of Quantitative MicroAnalysis (Milton. 7a Canongate Technology Ltd (1983) The diffusion of gases through a silicone rubber membrane. J. E (1955) Colorimetric analysis. A. C. L. and Snell.J.. Biol. J. E D. M. 31. pp. eds. J. 20 Vogel.M. Anal. II. (1975) Principles of Microbe and Cell Cultivation. Adv.

.M. (1981) Instrumentation for process control in cell culture. and Ribbons. New York. and Severinghaus. (1981) A porous membrane-carrier gas measurement system for dissolved gases and volatiles in fermentation systems. Proc. (1986) The electrochemistry of gases of medical interest. and Vezina. I. 22a. Kell.E. Methods 1. Goldstein. Dunn... J. L. Academic Press. J.J. J. J. Chromatogr. S. Calarco. D. 1. L.N. and Mfirdh. Ann. Peterson. 52 Perry. Untereker.. In: Gas Chromatography/Mass Spectrometry Applications in Microbiology (Odham. 32 van Stekelenburg. Chem.. S. John Wiley & Sons. Physiol. 54 Cramers.. R. Int. J. 48 Vurek..). D.S. Bacteriol. Clin. Scott. R. and Vos.I. Fed. 243-245. A. (1983) Gas concentration determination in fermentors with quadrupole mass spectrometer. 49 Coon. Trends Biotechnol. Biosensors 1.). L. (1958) A continuous infrared analyser for measurement of CO 2 in effluent air from bacterial cultures.. Silverman. R. Biochem.174 31 Rozzi. K. Soc..measurement of dissolved gases in fermentation liquids.G. Marcel Dekker. and Severinghaus. Norenberg.J. O. and other groups in aqueous solutions and tissue. M.F.. E. Ann.A. John Wiley & Sons. Feustel. 0985) Response of ion-selective field effect transistors to carbon dioxide and organic acids. I.W. J. (1984) Gas chromatography. (1976) Evaluation of a dual-function pH and pCO 2 in vivo sensor. Proc. (1983) Membrane inlet mass spectrometry . pp. P. E. 40.R. D. (1978) Infrared spectroscopy of ligands. Appl. 509-513. (1984) Direct bicarbonate determination in anaerobic digester liquors by measurement of the pressure of carbon dioxide.S. G. (1981) Introduction to Analytical Gas Chromatography. J. Exp. 41. 40 Lloyd. 34 Panteghini. 1995-1998. and Garland. I. and Larsson. Amsterdam. 48. In: Gas Chromatography/Mass Spectrometry Applications in Microbiology (Odham. 60-63.-A. Bolzern..-A. and van Wijngaarden-Penterman (1985) An extremely simple and fast microlitre method for the determination of total carbon dioxide in biological samples. (1983) Gas chromatography. D. and Labellarte. Langer. (1985) Direct potentiometric analysis of sodium.A.J.. J. G. C. Elsworth. R. C. T. R. Ion Phys..A. I. J. 50 Nicholls. Biol. M. J.V. P. Chromatogr.G. eds. W. R. H.. G. and Meyerhoff. J. 35 Odham. VaIL C.R.. Appl. 36 Larsson. and Bonora. Biomed. M. W. 22. and Buckhold.. J. Malchiodi.. 1483. and Williams.L. J... Biochem.. Clark. 55 Willett. L. eds. J. 4. London. eds. Eng.. Fitzgerald. Plenum Press. 1-6. Lai. C. C. pp. G. (1982) Fiber optic HCO 2 probe. pp. Methods Enzymol. G. 46 Peterson.E. 1. D. Chem. Anal. Larsson. A.L. and Albery. L. G. Rev. Libr.. New York. D. Weaver.C.G. Prod. Adv. S. 51 Fogt.B. Oxford. 42 Heinzle. 439-444. 313-328. Morris. R. Goldstein. A.W. Microbiol. J. (1977) Modern Practice of Gas Chromatography. In: Electrochemistry. J. 43 Tebbutt. 3. W.. D. 47 Vurek. J. and Mfirdh. Clin. 21.W. R. Elsevier. Proc. and total carbon dioxide in serum. 302-323. (1987) Gas Chromatography.. 44 Telling.C.. B. New York.). G. gases. Pergamon Press.. Appl. (1976) Enzymatic method for determination of CO 2 in serum. Boh~ttka.). (1980) Fiber optic pH probe for physiological use. (1985) Quadrupole mass spectrometry in the monitoring and control of fermentations. 277-280. 22. 54. R. P. Mass Spectrom. and Berecz. 625-629. 55-63. M. and Scott. A.G. D. eds. 45 Maxwell.R. 39 Lloyd. J. and Sinskey. G.. New York. Vol.. J.. Scientific and Engineering Principles (Moo-Young. potassium. eds.. M. 26-44.P. and McNair. 57 Clarke. J. Chem.C. J.. 27. J. D. Larsson. D. Jan.54. pp. 7-26.. Clin. Plenum Press. E. P. D. 179-212. (1972) Electrode measurements of carbon dioxide. R. N. In: Advances in Biotechnology. P. (1982) The role of ion-selective electrodes in .. 57. and Szil~gyi. Szilfigyi. M. In: Methods in Microbiology 6b (Norris.G. C. 2 7 .. L. Sensors and Analysis (Smyth. (1980) The role of ion-selective electrodes in improving fermentation yields. G. and Pierre. 38 Lloyd. and East. and Kampine. 56 Kell. 37 Fleischaker. K. pp. 864-869. Am. Robinson. 19. and Odham. G. (1983) Direct measurement of dissolved gases in microbiological systems using membrane inlet mass spectrometry. Fed. D. Microbiol. J. 201-203.V.). R. 499-510. Hahn.J. J. and Burns. and Caughey.W. 11. D. Int. A195-A224.N. B. 137-167. R. (1983) A fiber optic pCO 2 sensor. Biochem. 315-322. Robinson. (1984) Mass spectrometry. S. and Bourne. New York. 28-35. Anal.. 33 Forrester. 41 Bohfitka. I. 53 Grob. Wataji. chloride.

Amsterdam.S. Chem. London.. A. Shiokawa. T. and Collinge. eds. M. W. New York. J.J. 81 Eggins.175 microbial process control. Microbiol.. B. (1982) Electroanalytical Chemistry. 80 Bennett. S. J.. 77 Bond. (1987) Electrochemical techniques.A. 1 82.W. 64 Covington.S.M. eds. W. Bioeng. (1983) Photoassisted reduction of CO 2 over aqueous suspensions of strontium titanate powders and electrolytic reduction of CO 2 at titanate electrodes. 143-151. T. Eng. L.T. Enschede. 17. Neuman. S. 79-87. J.. D. J. 213-320. CRC Press. R..T. and Ko. (1984) Prediction of the dynamic response of the potentiometric carbon dioxide electrode.R. Soc. 356-359. and Symanski. (1976) Polarographic Oxygen Sensors.. (1982) Transcutaneous analysis of arterialpCO 2. 31.. Soc. 792-800.. E. Marcel Dekker. Community Ser.. 59 Donaldson. New York. Fleming.-M. 72 Fatt. R. and Barron. W. 71 Bruckenstein. D 2.. Acta 166. and Ingold. and Suzuki. T. B.H. In: Chemical Sensors (Seiyama. A. and Sawyer. (1986) Continuous conductometric sensor for carbon dioxide. pp. Chim. IE Trans. A. D. J. and Pirrwitz. Centrum voor Micro-Elektronica..C. Biomed. Electroanal.A.H. thesis. S.. A. Calder. and Biomedical Applications of Solid State Chemical Sensors.T. Palz. New York. Part I. and Severinghaus. pp. 316. Corrosion and Metal Finishing . Inst. D. 62 Gauthier. T. Shimizu. (Baltimore) 16. (1978) Theory. K. 22. K. Soc. (1984) Noninvasive transcutaneous monitoring of arterial blood gases.2"bipyridine). 2411-2416. and Kurachi. R.. Anal. London Ser. 332-338. R. P..H. Phil. J.C. and Bergveld. S. West Palm Beach.. and Palmer. and Sibbald. (1980) Modern Polarographic Methods in Analytical Chemistry. CRC Press.. Fueki. A general survey of voltammetry at different electrode materials in different solvents.. Photoelectrochemical and Photobiological Processes (Hall. In: Methods in Microbiology. (1985) The development and application of biosensing devices for bioreactor monitoring and control.L. Fujiwara. 1766-1770. W.V.R. and McNeill. Chem. J. Koster. John Wiley & Sons. D. A... and Symanski. eds. Amsterdam.. M. Einsele. Design. R. 58. J. D. 17-24.M. A.J. Faraday Trans. Lehn.K. 65 van den Berg. 75-131.. Sol. Eggins. D. Carr. (1982) A membrane electrode for the determination of CO 2 and 02. and Roberts. Instrumen. Elsevier. B. Med. D. 63 Cheung. 22. Vol. (1988) Ion Sensors Based on ISFET's with Synthetic Ionophores. E. In: Techniques in Electrochemistry. R. B. T. (1974) Experimental Electrochemistry for Chemists.). Chem. Proc. Energry R D Eur. Belanger. J. and Ribbons. 75 Albery. (1983) Solid carbonate electrolytes for the potentiometric measurement of carbon oxides and carbon bearing materials. W. A. J. 55-74.A Handbook (Kuhn A. T. Chem. Trans. M. Buhler. 93-101. and Suzuki.). London. 82 Tinnemans.).D. G.A. P. P. Yotsuya. and Mackor. 67 Hazinski.J. A. 60 van der Schoot..P. A. and Ziessel. In: Photochemical.. eds. Blake-Coleman. 353-356. Can J. I. 79 Plambeck. (1980) Steam-sterilisable pCO 2 electrode. John Wiley & Sons. K. D. T. ed. J. 138. B. pp. 83 Hawecker. Shiokawa. Moody. K. Y. R.. 4. W.. 58 Clarke. 585-590. Thewissen. K. J. 25.. 6b (Norris. Fueki.. 68 Hagihara. 39. Y.A. 25-53. J. 150-153.G. WileyInterscience. (1967) Electrochemistry of carbon dioxide in dimethyl sulfoxide at gold and mercury electrodes. J. (1987) Ion-selective field-effect transistors (ISFETs). (1976) Measurement of dissolved CO 2.T. Chem. (1984) Electrocatalytic reduction of carbon dioxide mediated by Re(bipy)(CO)3Cl (bipy=2.. Biosensors 1. and Ribbons.S.M. E. and McMullan. H. Comm. Anal.L. and Peura. 328-330. S. H. and Fauteux. Basic Principles and Applications. 66 Alford. 148. Ohkawa. 74 Haynes. D. Elsevier. Ph. 61 Puhar. pp. Chem. (1972) Oxygen electrode measurements. Academic Press. M. J. (1979) Dynamic response of the carbon dioxide electrode. A. Electroanal. 70 Bruckenstein. Am.. Eng. .R. Chem.). W. (1980) Recycling carbon dioxide from fossil fuel combustion. 1105-1116. Hasegawa. 73 Beechey. Anal. 69 Mendelson. Ion-Sel.W. New York. W. McNeill. 76 Sawyer. 31-46. (1983) Transcutaneous CO 2 electrode. (1986) Analytical applications of gas membrane electrodes. Chem.). In: Chemical Sensors (Seiyama. J. Electr. D. 78 Kuhn. J. S. J. A. (1983) Voltammetry of carbon dioxide. Rev. Anal. 52-56. B. Biotechnol.O. 86-91.

H. 103. Chiesi-Villa. (1987) Carbon Dioxide as a Source of Carbon. Chem. 61.A.. Yokoyama.213 -222. 88 Aresta. N. Proc. Aberystwyth. 746-748. and Lawrence. (1988) Application of microbiological sensors in fermentation processes.T.. Aurian-Blajeni. Fujishima. (1989) Piezoelectric crystal monitor for carbon dioxide in fermentation processes. Suzuki. Am. (1981) Physical Chemistry with Applications to Biological Systems.. (1987) Methanation and photo-methanation of carbon dioxide at room temperature and atmospheric pressure. C. M. 1323-1336. 373 . Soc.. U. Thewissen. Chem. C. and Ziessel. Natl. D. and Sawyer. I. (1985) Photoelectrochemicai reduction products of carbon dioxide at metal coated p-GaP photocathodes in aqueous electrolytes.D. (1982) Photoreduction of carbon dioxide to formic acid. T. Thesis. (1983) Efficient photochemical reduction of CO 2 to CO by visible light irradiation of systems containing Re(bipy)(CO)3X or Ru(bipy)32+ . Soc. H. M.. 536-538. Soc. 6.. 637-638. Koster. and Mackor. 101 Roberts. A. and Halmann. S. .. 100 Hawecker. 98 Tinnemans.379. Tamiya. and Iida.. 79.. Soc. R. A. 157.M. 87 Mills. O. J. 288-295. 1-7. University College of Wales.. Sci. (1982) Formation of methanol and other C1-C3 compounds in the photoassisted reaction of formaldehyde and water over strontium titanate suspensions containing transition metal oxide deposits. and Gratzel. Tamiya. Soc. J. Chem.. Am. E. 107. M. (1985) Determination of electroactive and non-electroactive gases using a membrane polarographic detector in a flow system. Lett. and Ziessel.. J. 90 Taniguchi. K. and Guilbault. J. Energy 1. 701-704. (1983) Photo-aided reduction of carbon dioxide to carbon monoxide. B. (1988) Carbon dioxide sensor using thermophilic bacteria. (1980) Electrocatalyic reduction of carbon dioxide by using macrocycles of nickel and cobalt. Anal. Ph.W. J. M. D.. I. S. formaldehyde.M. methanol. Yoshida. G. Int. J. E. 6278-6282. 89 lnoue. K.176 84 Chang.. T. I. Ohta.M. Kitagawa. 92 Ikeda.A. B. R. (1985) Stepwise reduction of carbon dioxide to formaldehyde and methanol: reactions of CO 2 and COs-like molecules with hydridochlorobis(cyclopentadienyl)zirconium (IV). and Mackor. Dordrecht. J. 85 Suzuki.. Soc. Nature (London) 327. (1984) On the reduction products of carbon dioxide at a p-type gallium phosphide photocathode in aqueous electrolytes. (1988) Effects of CO 2 on anaerobic bacterial growth and metabolism. R. J.A. Reidel Publishing Co. Y. J. 91 Ito.C o s+ combinations as homogenous catalysts. Comm. K. J. 7361-7363. Anal. Suleiman.. T... H.A. Electroanal. (1979) Photoelectrocatalytic reduction of carbon dioxide in aqueous suspensions of semiconductor powders. Mackor.. 99 Gambarotta. J.H. Aurian-Blajeni. (1984) Tetraazamacrocyclic cobalt (II) and nickel (II) complexes as electron-transfer agents in the photo(electro)chemical and electrochemical reduction of carbon dioxide. R. de Andrade. acetaldehyde and ethanol using suspensions of strontium titanate with transition metal additives. Neth. A. E. Nature (London) 277. (1982) Photochemical generation of carbon monoxide and hydrogen by reduction of carbon dioxide and water under visible light irradiation. Analyst 110. Ikeda.W. 97 Lehn. D. 9. O'M. 69-77. Bull. Tinnemans. Chem. 86 Karube. Lehn. D.-M. Kiwi. Acad. H. M. A. London. Anal. 23-26. A. Biochemical and Chemical Uses. Chim. Jpn. 102. Acta 213. Y. 1353-1357. J.. Chem. and Bockris. L. K. and Oshima.H. and Forti. 58.S.G. Chem. B. A. 102 Dixon. and Honda. 95 Thampi. Chem. K. 506-508. Konishi. A. Thewissen. S. J. G. A. 94 Ulman. Chem.. and Eisenberg. T. Nouv. (1965) Voltammetric determination of carbon dioxide using dimethylsulfoxide as a solvent. 93 Tinnemans.-M.. 57. Electroanal. 21.P. 583-584.A.. C. 96 Fisher. M. K.. Chem. Chim. J. Karube. R. Collier Macmillan Publishers. Koster. S.. Jpn. Bull. Chem. S. and Asano. H. Floriani. T. and Ito. Sol. P.. R. 179-182. Sode. M. S. Strologo. and Guastini. A... A. 103 Fatillo-Filho.