You are on page 1of 5

PII: S0043-1354(99)00217-1

Wat. Res. Vol. 34, No. 3, pp. 1037±1041, 2000 # 2000 Elsevier Science Ltd. All rights reserved Printed in Great Britain 0043-1354/00/$ - see front matter

Department of Environmental Technology, Wageningen Agricultural University, Bomenweg 2, 6703 HDWageningen, The Netherlands (First received 1 October 1998; accepted in revised form 1 May 1999) AbstractÐThe in¯uence of temperature on the oxygen transfer rate (OTR) was studied in a bubble column. Aeration took place in three di€erent liquids: tapwater, anaerobically pretreated paper process water and thermophilic sludge grown on a mineral medium and volatile fatty acids as carbon source. The OTR was measured in a temperature range of 20±558C in case of tap- and process water. The OTR in the thermophilic sludge was determined at 558C.The OTR remained constant over the speci®ed temperature range in case of tapwater and showed a slight increase in case of process water. The constant OTR in case of tapwater was due to the counteracting e€ect of an increased overall oxygen transfer coecient versus the decreased oxygen saturation concentration at higher temperatures. At 558C the OTR in the thermophilic sludge was comparable to both other liquids at this temperature. # 2000 Elsevier Science Ltd. All rights reserved Key wordsÐoxygen transfer, Kla, forest industry, wastewater, thermophilic, aerobic



y a c

theta factor, typical value approximately 1.024 (±) interfacial surface area (m2 mÀ3) dissolved oxygen concentration in the bulk liquid (mg O2 lÀ1) ce apparent oxygen saturation concentration, observed in the experiment (mg O2 lÀ1) cs saturation concentration of dissolved oxygen at the speci®ed temperature, pressure and salinity (mg O2 lÀ1) bubble diameter (m) db oxygen di€usion coecient in the liquid phase Dol (m2 sÀ1) oxygen transfer coecient (m hÀ1) Kl Kla(T ) overall oxygen transfer coecient at temperature T (hÀ1) OC oxygenation capacity (mg O2 lÀ1 hÀ1). OTR oxygen transfer rate (mg O2 lÀ1 hÀ1) R gas constant (=8.314) (J molÀ1 KÀ1) ract actual respiration rate of the biomass (mg O2 lÀ1 hÀ1) t time (s) T temperature (8C in equation (2), K in equation response time of the oxygen probe (s) bubble rise velocity relative to the liquid (m sÀ1)

tp vbs


*Author to whom all correspondence should be addressed. Tel.: +31-317-484-993; fax: +31-317-482-108; e-mail:

The paper and board industry is putting a lot of e€ort in closing the watersystems in the production line resulting in the so-called zero discharge paper mills. Operation of these zero-discharge mills was shown to be possible in the board industry but requires an in-line treatment system to prevent bad smells in the endproducts. Recently, Habets and Knelissen (1997) presented a sequenced anaerobic± aerobic treatment system as the most cost-e€ective option for in-line treatment of process water from a paper mill using recycled wastepaper as raw material. A disadvantage of their set-up is the required cooling of the process water to mesophilic conditions prior to biological treatment and subsequent heating afterwards. The process water needs to be reheated since the optimal temperature for paper production is approximately 558C. The current research project aims at a complete anaerobic± aerobic treatment under thermophilic conditions (50±558C) of paper process water in a closed cycle mill. Regarding the post treatment step, it is believed that thermophilic aerobic treatment may be limited by a poor oxygen transfer due to the lower oxygen saturation concentrations (cs) at higher temperatures. The goal of this study is to determine the OTR under mesophilic and thermophilic conditions in tapwater and process water.


vapour pressure corrected. Bewtra et al. 1993) with volatile fatty acids as organic substrates. The thermophilic sludge was substrate depleted to assure a constant respiration rate during the Kla determination. ract ˆ Kl a…cs À ce † …3† The oxygenation capacity (OC) of the aeration system is calculated as the OTR assuming a zero oxygen concentration in the bulk liquid. The oxygen saturation concentration (cs) in the thermophilic sludge was estimated by equation (3) which is similar to equation (1) when a steady state during aeration is reached. The respiration rate was measured by switching o€ the aeration while stirring was maintained.09 7. The probe response time. (1970) measured the OTR in a 40 m3 aeration tank and found a constant OTR over a temperature range of 10±308C. vapour pressure corrected. At higher temperatures the OTR was often estimated by use of the theta factor (y ).50 6. d Tapwater.15 9.3. (1990) found that the OTR showed only minor changes in a temperature range between 15 and 708C in a mechanically mixed fermentor while in an air-mixed fermentor the transfer capacity for O2 decreased slowly but steadily. Boogerd et al. The OTR was determined in a temperature range of 20±558C in case of tapwater and process water while the OTR in the sludge was measured at 558C. i. These e€ects result in an increased Kla value that might o€set the smaller driving force. In all three experiments the Kla values were determined in duplicate with the `start-up dynamic method' as described by Linek et al.e. Materials and methods The experiments were performed in three di€erent liquids: tapwater. surfactants. The theta factor is de®ned by equation (2) and can be used to estimate Kla values at higher temperatures in case the value at 208C is known.10 5.15. this studyd 9.43 6.2 7. An overview of the oxygen solubility data from various sources. The thermophilic sludge in experiment 3 was grown on a mineral medium (Schlegel. Relation between OTR and temperature Increasing temperatures result in a lower oxygen solubility leading to a smaller driving force (csÀc ) and hence to a lower OTR. A cylindrical bubble column with an e€ective volume of 3 l was aerated at di€erent ¯ow rates. (1990)c 9. 0. with and without correction for the vapour pressure are presented in Table 1. not vapour pressure corrected. Stenstrom and Gilbert (1981) note that the use of The use of an autoclavable stainless steel oxygen probe can lead to additional diculties because of the electrode response time. 0.45 and 0.67 3. c x5 mM sulphuric acid. very di€erent values were found for this factor under well de®ned experimental conditions.56 vvm (volume air volume liquidÀ1 minÀ1). tp is de®ned as the time needed to record 63% of a stepwise change. Stenstrom and Gilbert (1981) mention the following: air ¯ow rate bubble diameter temperature viscosity basin geometry (a€ects contact time between gas and liquid) wastewater composition (salts. anaerobically pretreated paper process water and thermophilic sludge. A large tp can lead to an inaccur- Table 1.65 Boogerd et al.41 5. Mixing took place by aeration and a magnetic stirring device. The column was temperature controlled.84 2.5 5.41 5.50 5. dc ˆ OTR ˆ Kl a…cs À c† dt …1† this factor is inadequate to predict oxygen transfer at higher temperatures. In experiment 2 the process water was aerated for a short period to oxidise possible interfering compounds (sulphides) before the Kla determination took place.43 7. the di€usion rate of oxygen increases with increasing temperatures while the liquid viscosity and surface tension decrease. vapour pressure corrected.09 7.98 7.56 6.82 5. The response time (tp) of this electrode is larger than the conventional oxygen probes that are used under moderate temperature conditions.1038 Research Note The oxygen transfer rate is in¯uenced by several factors. Kl a…T † ˆ Kl a20 yTÀ20 …2† However. 1924) describes the rate of oxygen transfer from a gas to a liquid phase. Demineralized water.1 Demineralized water. Oxygen solubility data (mg lÀ1) in aqueous solutions at p=p0 Temperature (8C) APHA (1995)a 20 30 40 50 55 60 70 80 a b Oxygen solubility Perry and Green (1984)b 9. Oxygen was measured with an autoclavable Mettler Toledo Inpro 6000 oxygen probe (silicon membrane) connected to a Knick 7302-2 oxygen meter.11 6. .07 4. However. 0. The set-up is presented in Table 2.55 6.88 Measured. biomass) Equation (1) (Lewis and Whitman. (1993).

RESULTS AND DISCUSSION The results of the aeration experiments in tapwater are presented in Fig. This is mainly due to a lower cs in the sludge compared to the tapwater. Kla values determined at 558C in tapwater (W).3 vvm. The Kla increases as expected while cs decreases in the same extent. Standard errors were less than 1%. 1984). The OC in the tapwater remained constant in the temperature range of 20±558C while the OC in the process water increased. Several correction programs were designed in the past to compensate for the dynamic behaviour of the probe (Aiba et al. Fig. the solubility data from Perry and Green (1984) are not corrected for the increasing water vapour pressure at higher temperatures and overestimate the actual saturation concentrations. 2. Smith and Skidmore (1990) determined gas hold-ups and Kla values in a concentric tube airlift reactor at 30. Van 't Riet (1979) states that it is possible to measure Kla values without additional correction programs if equation (4) is met. 1984. The OC in the thermophilic sludge at 558C is comparable to the OC in the process water and slightly lower compared to tapwater. Comparable results were found by other researchers. Boogerd et al. tp‚ 1 3Kl a …4† The above mentioned restriction was always met in this research and no additional correction programs were used. 2. The e€ect of the air ¯ow rate on the Kla is shown in Fig.Research Note Table 2. Calculation of the OTR with their Kla data and oxygen saturation concentrations from Boogerd et al. (1996) also found increased oxygen transfer rates with increasing temperatures in a mechanically mixed fermentor. Wise and Houghton (1966) give the following relation: Dol ˆ 4X2  10À6 exp…À18  10À3 aRT † …5† For bubbles larger than 2. results in a slightly increasing oxygen transfer rate with increasing temperatures. Krahe et al. 50 and 728C. This results in a higher Kl value thus having a positive e€ect on the OTR. The cs values that were measured (exp.56 vvm. These saturation data were used to calculate the oxygenation capacity (OC) of the system in the di€erent media at the various temperatures (Fig.15 and 0. 3). The Kla values in the sludge and the process water do not di€er signi®cantly from the values found for tapwater. 1. With regard to the di€usion rate for oxygen. process water (q) and thermophilic sludge (r) at air ¯ow rates between 0. 1 and 2) or estimated (exp. cs (r) and Kla (Q) values measured in tapwater with a constant air ¯ow rate of 0. 3) are presented in Table 3. 1984 Fig. However. The Kla values and hence the OC are linearly related to the air ¯ow rate within the speci®ed range. Experimental set-up Liquid Experiment 1 tapwater Temperature (8C) 20 30 40 55 20 55 55 1039 Experiment 2 Experiment 3 process water thermophilic sludge ate calculation of the Kla if the dissolved oxygen concentration changes rapidly in time. Koizumi and Aiba. the bubble characteristics remain unchanged in this ¯ow range. Apparently. The increased temperature will lead to a higher di€usion rate for oxygen (Dol) in the liquid phase due to the direct temperature e€ect and indirect by the lower liquid viscosity.. in some extent predictable by the general mass transfer theories. (1990). . The present results are.5  10À3 m the bubble surface is always mobile and Kl can be calculated using the Higbie model Heijnen and Van 't Riet. (1990) warns for this e€ect since it also leads to an overestimation of the OTR. They determined the Kla values experimentally and used oxygen solubility data from Perry and Green (1984) to calculate the OTR. especially at higher temperatures. 1.

Bioeng. Koizumi J. N.. K. So. Air ¯ow rate 0. Bewtra J. the bubbles generated from a porous disc/stone are approximately (2±4)  10À3 m and tend to coalesce to an ultimate bubble diameter of 6  10À3 m (Heijnen and Van 't Riet.2 38. 3. in the worst case. G. M. This is in accordance with the theory for tapwater.. (1984) The e€ect of temperature on Kla in thermophilic cultivation of Bacillus Stearothermophilus.8 cs in thermophilic sludge was estimated by equation (3).1 30. H. Washington. (1970) E€ect of temperature on oxygen transfer in water. G. and Van der Lans R. Shi Ru J. The most optimal case predicts no temperature e€ect on the interfacial surface area. C. Biotechnol. and Kappa Packaging.8 21. J. In a coalescing medium.  Kl ˆ 4Dl vbs pdb 0X5 …6† Visual inspection showed that the bubble diameter in this research was larger than 2.8 Kla process water 17. process water (q) and thermophilic sludge (r) versus temperature. (1982) found no e€ect of liquid viscosity on the gas hold-up.6 20. REFERENCES Fig. 26.V. (tapwater).0 20. and Knelissen H. and Polkowski L. Heijnen J. Bos P. 1136±1138. According to Stokes law the terminal bubble rise velocity in case of rigid spheres (small bubbles) is inversely proportional to the liquid viscosity.7 21. APHA (1995) Standard Methods for the Examination of Water and Wastewater.. Bioeng. Kuenen J.4 20. However. 14. Res.5 5. DC. R. Senter EET project 96003. The temperature e€ect on the separate variables Kl and a has not been revealed yet.5  10À3 m.42 1. The latter seems more likely when looking at the results presently found. Bioeng. This e€ect might be slightly positive. negative or none at all (this research) depending on the bubble diameter and the reactor type. when Stokes law applies a 50% decrease in the interfacial surface area can be expected with a temperature increase from 20 to 558C. The oxygen saturation concentration decreases 44% with a temperature increase from 20 to 558C (this research).55 39.43 6. Habets L. (1997) In line bio- . Biotechnol. Wat. This leads to a lower gas hold-up and hence a smaller interfacial surface area. Biotechnol. 19th ed. American Public Health Association.4 26.19 7.5 37. 35. J.3 vvm. CONCLUSIONS The present research has shown that in tapwater at higher temperatures the decreasing oxygen saturation concentration (cs) was completely o€set by the increased overall oxygen transfer coecient (Kla ).. (1982) Oxygen mass transfer into aerated CMC solutions in a bubble column. possible changes in interfacial surface area are not taken into account yet. Aiba S. autotrophic cultivation of moderate and extreme thermophiles: a case study related to the microbial desulfurization of coal. D. Deckwer W. Boogerd F. B. and Mukhopadhyay S.3 vvm. Nicholas W. and Schumpe A. In general it can be concluded that the oxygen transfer rate is only slightly a€ected by the liquid temperature. The research is ®nanced by the Dutch Ministry of Economic A€airs. 461±481. Nguyen-Tien K. 115±123. OC in tapwater (W).5 cs process water 8.15 Kla tapwater 22. 1984). J. 4. Equations (5) and (6) predict a 48% increase in Kl in the same temperature range thus completely counteracting the lower saturation values.6 cs thermophilic sludge Kla thermophilic sludgea 4. In case of bigger bubbles (1 cm) Deckwer et al. This results in a constant oxygen transfer rate in tapwater within the temperature range of 20±558C.7 4.. It is not expected that the bubble diameter will change with temperature but a lower liquid viscosity will most probably result in a higher bubble rise velocity. AcknowledgementsÐWageningen Agricultural University is co-operating in this project with Paques Water Systems B. (1990) Oxygen and carbondioxide mass transfer and the aerobic.1040 Research Note Table 3. A. Measured cs (mg lÀ1) and Kla (hÀ1) values in tapwater and process water Temperature (8C) 20 30 40 55 a cs tapwater 9. Air ¯ow rate 0. The Kla values in these liquids do not di€er signi®cantly from the values found for tapwater. The slightly lower OTR values in the process water and the thermophilic sludge compared to tapwater are caused by the lower oxygen saturation concentrations in these liquids.42 0. 1111±1119...

(1993) General Microbiology. and Sinkule J. Perry and Green (1984) Perry's Chemical Engineers Handbook. Biotechnol. (1993) Measurement of Kla by dynamic pressure method in pilot-plant fermentor. Biotechnol.. FEMS Microbiol. L. Wat. Cambridge. Ind. 357± 364. . Smith B. and Whitman W. Bioeng. 271±285. Moucha T. 1215. 643±654. (1979) Review of measuring methods and results in nonviscous gas±liquid mass transfer in stirred vessels. 35. 16. G. Eng. Heijnen J. Lewis W. Eng. 26. J. 43. (1966) The di€usion coecients of ten slightly soluble gases in water at 10±608C. K. 999±1010. and Van 't Riet K. Krahe M. 483±491. speci®cation and operation of aeration systems. and Gilbert R. Antranikian G. K. Linek V. Dev. and Houghton G. Chem. 18. Bioeng. 15. 18. B21±B42. Wat. 41±48. New York. and Aiba S.. J. 1131±1133. R. 1041 Schlegel H. Koizumi J. Dousova M. Process Des.. G. Technol. McGraw-Hill. Sci. Eng. (1990) Mass transfer phenomena in an airlift reactor: e€ects of solids loading and temperature. G. Chem. beta. Res. (1981) E€ects of alpha. and Skidmore D. (1996) Fermentation of extremophilic microorganisms. C. Sci. 35. (1984) Reassessment of the dynamic Kla method. and theta factor upon the design. Stenstrom M. Chem. University Press. Chem. (1924) Principles of gas absorption. Van 't Riet K. 28. Bioeng. Eng.Research Note logical water regeneration in a zero discharge recycle paper mill. Biotechnol. 21. (1984) Mass transfer. Ind. Rev. Wise D. and Markl H. mixing and heat transfer phenomena in low viscosity bubble column reactors. 477±482.