- Email: [email protected]
A chemical and sensory study of odour compounds in the Athabasca River, Alberta, Canada
e>
Pergamon
Welt. Sci T.ch. Vol. 31. No. 11. pp. 1S-21.
Copyright C
199~
199~.
IAWQ
Printed iD Great BritaiD. All rigbll reserved.
0273-1223(95)00450-5
0273-1223,o}~ S9'~0 +
0-00
A CHEMICAL AND SENSORY STUDY OF ODOUR COMPOUNDS IN THE ATHABASCA RIVER, ALBERTA, CANADA S. L. Kenefick*, B. G. Brownlee**, T. R. Perley* and S. E. Hrudey* • Environmental Health Program. University ofAlberta, Edmonton. AB. T6G 2G3. Canada •• National Water Research Institute. 867 Lakeshore Road, Burlington. ON. L7R 4A6 Canada
ABSTRACT As of early 1993. the Athabasca River received effiuent from one bleached kraft pulp mill (Mill A). three cbemithermomcebanical pulp and paper mills. one oil sands extraction and upgrading plant and a number of municipal effiuents. In the latter half of 1993 a second bleached kraft pulp mill (Mill B) began operation midway along the river. An investigation was carried out to cbaracterize the odours in the river water using both chemical and sensory methods. in a pre- and post-operational study of the second bleacbed kraft mill. Both surveys were earned out under ice during low flow conditions. Samples were analysed by gas cbromatograpby-mass spectrometry after extraction using a closed-loop stripping apparatus (CLSA). In addition, sensory analysis by flavour profile panel and olfactory gas chromatography of the CLSA extracts were performed. In 1993 all analytical methods conftrmed that compounds characteristic of bleached kraft mill effiuent were deteCtable for more than 950 km downstream from Mill A and that this effiuent was the major source of odour to the Athabasca River. Cbemical and sensory results for 1994 samples (collected after Mill B began operation) indicated a decrease in the impact of Mill A compared with the 1993 survey. The Mill B effiuent bad distinctive odour but its impact on the river was difficult to detect due to dilutioo and background odour from the Mill A effiuent.
KEYWORDS Bleached kraft pulp mill; chemithermomechanical pulp mill: closed-loop stripping; flavour proftle analysis; odour; olfactory gas chromatography
INfRODUcnON Bleached kraft pulp mill effluents are known to contain odorous compounds that have the potential to cause taste and odour problems for downstream users (Cook et aI., 1973; Kovacs and Voss, 1986; Paasivirta t't aI., 1910; Wigilius t't al., 1988; Wong et al., 1985). Such effluent has been reported to impair the taste and odour of drinking water at eft1uent concentrations ranging from 0.1 to 0.4% (Kovacs and Voss, 1986). Although most of the compound-specific work on pulp mill effluent odours has focused on chlorinated compounds, the literature does not provide a reliable basis for attributing odour problems primarily to 15
S. L. KENEFICK et al.
16
chlorinated compounds. Work by Domw Fine Papers (1971) to identify the most odorous process streams did not identify specific compounds but the process streams identified were rich in odorous organosulphur compounds. The production of these compounds is independent of bleaching practices and their control depends on good in-plant spill control and efficient wastewater treatment. Headley (1987) found odorous suiphones, sulphides and thiophenes in pulp mill effluents. He reported that sulphones and sulphides originated during biological treatment of kraft mill waste streams and thiophenes were formed as a result of the lignin digestion process, in which sodium sulphide and sodium hydroxide are used. Pulp mill effluent contributions to river water levels of phenols and trichlorophenol are also considered important after recent Swedish reports of bio-methylation of trichlorophenol to 2,4.6-trichloroanisole, an extremely potent source of musty odours in water and fish (Nystrom et ai., 1992). In early 1993, the Athabasca River received effluent from one bleached kraft pulp mill (Mill A), three chemithermomechanical pulp and paper mills, one oil sands extraction and upgrading plant and a number of municipal effluents. In the latter half of 1993 a second bleached kraft mill (Mill B) began operation midway along the river. A more complete description of the study area can be found in Brownlee et ai. (1995). Preliminary studies in 1991 had shown that odour compounds including 2,4,6-trichloroanisole and several chlorinated veratroles persisted for more than 1000 km downstream of Mill A (Brownlee et ai., 1993). Our investigation was carried out to further characterize the odours in the river water using both chemical and sensory methods. in a pre- and post-operational study of the second bleached kraft mill. The pre-operational survey was conducted in February and March of 1993 (Kenefick et ai., 1994) and the second or "post• operational" phase was conducted in February and March of 1994 (Kenefick et ai., 1995). This second phase of the study was also carried out after the condensate recovery system at Mill A was upgraded and its bleaching process was changed from 45% to 100% chlorine dioxide substitution. The primary objectives of this study were: I) Carry out comprehensive chemical, olfactory gas chromatographic. and flavour profile analyses to describe odours and their persistence in the Athabasca River; 2) Carry out this study before and after Mill B begins operation to determine the impact of the second mill; 3) Evaluate the usefulness of each analytical method (chemical, olfactory, flavour profile) for this kind of study and the overall value of using all three methods. METHODS The materials collected for both phases of this study included industrial effluent, municipal effluent, tributary, mainstream river and treated water samples. All water and effluent samples were collected at the time-of-travel of the Athabasca River as in Brownlee et ai. (1995). All were collected with no headspace, without addition of a preservative, and were cooled until delivery to the laboratory within 24 hours of collection. An 8 I water sample was collected at each river. tributary, and treated water sampling point and 2 I samples of each effluent were collected. E1ayour profile analysis The basics of the Flavour Profile Analysis method have been listed in a number of sources (ASTM, 1968; Bartels, et ai., 19H7; Krasner, 1988; Suffet, et ai., 1988) and these have been standardized in standard method #2170 (APHA-AWWA-WEF, 1992). The panel used in this study was run in accordance with this method. Closed-JoQp strippin~ and
~as
chromatompby-mass spectromet[y
One litre replicate subsamples of each river, tributary and drinking water sample were analysed, as received. All effluent samples were diluted 20:1 with odour-free water prior to extraction of 11 subsamples. A known
A chemical and sensory study of odour compounds
17
mass of I-chlorodecane was added to each sample, as a recovery standard, prior to CLSA. The samples were closed-loop stripped onto 1.5 mg carbon filters for 2 hours (water bath temperature at 30°C and filter temperature at 50°C) using a Brechbiihler AG closed-loop stripping apparatus. The filters were then extracted twice with 10 III and once with 5 III of carbon disulfide:acetone (9: 1). A known mass of biphenyl• d lO was added to the extract as an injection standard. A 2 III sample was analysed by GC-MS using injection in the splitless mode and 10 III of doubly distilled isooctane was added prior to subsequent analyses by OGC and GCIFIO. The GC-MS analysis was performed in the selected ion monitoring (SIM) mode on an HP 5890 GC with an HP 5970 mass-selective detector and an HP 59940 Chemstation® data system. In 1993 the GC-MS conditions were: 80°C to 180°C at 4°C/min and then lQoC/min to 280°C; 08-1301 column, 30 m x 0.25 mm ID, 0.25 11m film thickness; column head pressure 100 kPa; injector temperature 250°C; detector temperature 300°C. In 1994, with the added organo-sulphur target compounds, oven program and injection conditions were: 35°C to 90°C at 4°C/min and then 90°C to 280°C at lQoC/min; cool-on-column injector at 35°C and column head pressure 70 kPa. OIfactoQ'
~as
chromatoernphy
The CLSA extracts were also analysed by olfactory gas chromatography (OGC) using a Hewlett Packard 51190 GC where the conditions used in 1993 were: 80°C to 280°C at 4°C/min; 08-1301 column, 30m x O.32mm 10, 0.25 11m film thickness; column head pressure 70 kPa; injector temperature 250°C. In 1994, oven program and injection conditions were modified to: 35°C to 90°C at 4°C/min and then 90°C to 280°C at 10°C/min; with a cool-on-column injector temperature of 35°C and column head pressure of 70 kPa. For OGC, the column was raised through a heated transfer unit from the GC oven to a glass detection cone (olfactory detector outlet, SGE International). A 2 III sample was injected and after the solvent peak had eluted, the outlet was continually monitored over time. Elution time, intensity and odour descriptor were recorded. A six point intensity scale was used (1 = very weak, 6 = very strong). Control samples of known standard concentrations were used to help establish consistency in descriptors and intensity values. Method blanks and filter blanks were also analysed. RESULTS ANO DISCUSSION A summary of FPA, OGC and GC-MS results for a number of river sites are provided in Tables I and 2. In 1993 the FPA results demonstrated a strong increase in odour for Athabasca River samples downstream of Mill A when compared with the upstream sample. Panelists used descriptors similar to those used for the Mill A effluent and some panelists recognized these samples as reminiscent of pulp mill odour. There was little indication of any reduced intensity of odour until 1200 km downstream from the mill discharge. Other effluent samples (non-pulp mill) were only slightly more odorous than the tributaries and all were less odorous than the Mill A effluent. The flavour panel results indicate that the Mill A effluent was most likely the dominant source of odour in the mainstream river. These findings are consistent with the work of Kovacs and Voss (1986) wherein they found that biologically treated bleached kraft mill effluent could impair drinking water odour (as perceived by a flavour panel) at effluent dilutions in river water as high as 300 to 1000 fold. Flavour panel members also found that odour in the Fort McMurray (950 km downstream) water supply decreased substantially from raw to treated and descriptors that had been primarily associated with the Athabasca River samples downstream of Mill A were not used for the treated water. The oil sands upgrading plant effluent (downstream of Fort McMurray) was recognized as very distinctive and relatively strong, but with the possible exception of a sulphur descriptor it was difficult to recognize any impact of this effluent on the Athabasca River after dilution. The treated water at Fort Chipeweyan (1200 km downstream) had a very strong chlorine odour that likely masked any other subtle odours that may have been present.
S. L. KENEFICK et a/.
18
Table I. Distribution of odour in the Athabasca River (February and March. 1993) Distance Downstream from Mill A (km)
FPAAverage Intensity'" (most common descriptor)
OGCpeaks consistently detected.... (average intensity)
Target Compounds detected by GC-MS (Concentration in ngIL)
upstream
0.4 (odorless)
IBMP(1.5) 246TCA(1.5)
Mill A BKME
2.1 (sewage)
IPMP(2) sulphur(3) 246TCA(4.5)
20
1.4 (sewage)
IPMP(4) sulphur(2.5) 246TCA(3.5)
IPMP(3) 246TCA(l) 345TCY(4)
240
1.5 (pulp & paper)
IPMP(2) 246TCA(4) geosmin(2)
IPMP(l) 246TCA(3) geosmin(2) 345TCY(33)
400
1.8 (pulp & paper)
IPMP(2) 246TCA(4)
IPMP(2) 246TCA(4) geosmin(3) 345TCY(27)
550
1.2 (pulp & paper)
IPMP(3) 246TCA(3)
IPMP(I) 246TCA(6) geosmin(2) 345TCY(27)
650
1.6 (septic)
IPMP(2) 246TCA(2.5) geosmin(1.5)
246TCA(7) geosmin(2) 345TCV(24)
950
1.4 (pulp & paper)
IPMP(1.5) 246TCA(3)
246TCA(13) geosmin(2) 345TCY(26)
1100
1.2 (earthy)
IPMP(4) 246TCA(4) geosmin(2.5)
246TCA(5) geosmin(2) 345TCY(7)
1200
0.5 (earthy)
IPMP(2)
NID 246TCA(4) 345TCV(27)
NID
'wuh a scale maxImum of 3.0 ··Consistently detected = average intensity greater than I on a 6 point scale 246TCA = 2.4,6-trichloroanisole IPMP = 2-isopropyl-3-methoxy pyrazine IBMP = 2-isobutyl-3-methoxy pyrazine 345TCV = 3.4,5-trichloroveratrole Flavour profile result~ were similar in 1994 except that Mill A effluent odour was less intense and the odour descriptors changed to include woody and resinous. Mill B effluent had a similar odour but could not be differentiated in the river due to dilution and the background odour from Mill A effluent. The CLSNGC-MS analytical results demonstrated a number of interesting findings. The mainstream Athabasca River sample upstream of Mill A did not contain detectable levels of any of the target odour compounds other than geosmin (a biogenic compound) in 1993 or 1994. In the 1993 Athabasca River samples collected downstream of Mill A there was consistent detection of 3,4,S-trichloroveratrole (3,4,5• TCY) as far as 1100 km downstream. Geosmin and 2,4.6-trichloroanisole (2,4,6-TCA) were also detected in most mainstream samples downstream of Mill A (including Fort McMurray raw and finished waters). In 1994 the mainstream river samples collected downstream of Mill A did not show detectable levels of the target compounds that were found in the 1993 study. However, given the sulphurous odours detected using OGC in 1993. the 1994 GC·MS list of target compounds was expanded to include a number of thiophenes and thioethers (organosulphides). Low levels of thiophene and/or dimethyl disulphide (DMDS) were found as far as 750 km downstream of Mill A. This may have been caused by the presence of sewage treatment plant effluent in the Mill A discharge. The other three sewage treatment plant effluents all contained
A chemical and sensory study of odour compounds
19
thiophene or OMOS. However, the three biologically treated CTMP effluent samples also contained low levels of these sulphur compounds, indicating that a variety of biological treatment processes can result in their formation. CLSA analysis for target compounds did not provide evidence of causative agents which result in "pulp mill" odours. Table 2. Distribution of odour in the Athabasca River (February and March, 1994) Distance Downstream from Mill A (km)
FPA Average
Intensity· (most common descriptor)
OGe peaks consistently detected·· (average intensity) lPMP(2) IBMP(2)
Target COmpounds detected by GC-MS (Concentration in ngIL)
upstream
0.4 (odorless)
MillABKME
1.3 (septic)
NID
20
1.5 (septic)
sulphur(3) lPMP(3) IBMP(2) 246TCA(2)
240
1.3 (septic)
lPMP(3) IBMP(4)
thiophene(2) 01105(9)
400
1.3 (septic)
IBMP(2)
thiophene(1 ) geosmin(8)
550
1.3 (septic)
IPMP(2)
01105(2)
MillBBKME
1.4 (rubber/chemical)
650
1.4 (septic)
MIB(4)
750
1.5 (woody)
N/D
950
1.6 (septic)
IBMP(3) MIB(2) geosmin(2)
N/D
1100
1.0 (septic)
lPMP(2) IBMP(3)
N/D
1200
1.1 (septic)
N/D
N/D
sulphur(2) lPMP(2) MIB(4)
geosmin(9)
NID 01105(28)
NID O11OS(9) thiophene(3) 01105(12)
'WJih a scale maxImum of 3.0 ··deteeted = Intensity greater than 1 on a 6 point scale 246TCA = 2.4,6-trichloroanisole IPMP = 2-isopropyl-3-mcthoxy pyrazinc IBMP = 2-isobutyl-3-mcthoxy pyrazinc 34STCV 3.4,S-trichloroveralrolc MIB = 2-mcthylisobomcol DMDS = dimethyl disulphide
=
In the initial year of this study the predominant odours by oac in Mill A effluent were the 2.isopropyl-3• methoxy pyrazine (IPMP), sulphurous, and 2.4,6·TCA odours. One analyst also detected woody, sewage and spicy odours. The chemithermomechanical pulp mill effluents were generally low in odour but had some distinctive odour peaks described as woody, cardboard and sulphurous. The municipal wastewater treatment plant effluents also had low intensities but had a variety of distinctive odour peaks not found in the other samples, such as waxy, flowery, soapy, woolly, spicy and cucumber. In 1993 mainstream Athabasca samples (including the site upstream from Mill A), oac showed moderate to strong levels of 2.4,6-TCA and IPMP for 1100 km downstream of Mill A. Two unidentified musty cork odour peaks also showed up quite consistently at weak to moderate intensities. Geosmin occurred quite consistently at trace to moderate levels
S. L. KENEFICK et al.
20
and 3.4.5-TCV showed up occasionally at average intensities less than 1. Sulphurous and hydrocarbon odour peaks were also frequently noted. being strongest in the upper reaches and diminishing to non-detectable levels by Fort McMurray. The 1994 OGC results for the Mill A effluent differed sharply with those obtained for the same effluent in the 1993 study. The major process changes. including greater condensate recovery and 100% chlorine dioxide bleaching at Mill A, appear to have led to a significant reduction in odour peaks detectable with olfactory GC. The Mill B effluent sample showed weak levels of a sulphurous odour, and an IPMP odour and moderate levels of the camphorous odour of 2-methylisoborneol (MIB). In 1994 the mainstream Athabasca samples. including the site upstream from Mill A, showed weak to moderate levels of the musty pyrazine (lPMP or IBMP) odours as far as 1100 kIn downstream of Mill A. The musty. camphorous odour of MIB was detected in mainstream samples downstream of Mill B and as far downstream as Fort McMurray. The finished water sample collected at Fort McMurray had no detectable odour peaks and the Fort Chipewyan treated water had only a 2.4,6-TCA odour peak. CONCLUSIONS I) In 1993. and again in 1994, Kraft Mill A effluent was the major source of odour to the river. Low temperatures and ice cover allowed the odour to be transported long distances. This odour was detected at distances greater than 950 kIn downstream from the source. but the odour disappeared upon conventional water treatment, possibly due in part to masking by chlorine.
2) Mill B effluent did have a distinctive odour but its impact on the river was difficult to detect due to dilution and background odour from the Mill A effluent. CLSA analysis for target compounds did not provide evidence of causative agents for either effluent and a further expansion of the target compound list is required before conclusive. compound-specific tracing studies will be possible. 3) This paper summarizes olfactory GC, CLSNGC-MS and FPA results. but the three methods were difficult to correlate. The variety and intensities of odour peaks reported in OGC analyses gave a semi• quantitative statement about spatial distribution of odours. Flavour panel results gave an overall odour intensity for each sample and probably yielded the most useful information, but these FPA results could not be compared directly with chromatographic results. Unfortunately the current literature list of compounds responsible for the characteristic "pulp mill" odour was not applicable to the specific effluents being surveyed in this study. Unless the target compound list is redeveloped for each specific effluent, the three methods are difficult to link and they will remain difficult to correlate. ACKNOWLEDGEMENTS Funding for this research was provided by the Northern River Basins Study and the Natural Sciences and Engineering Council of Canada. The flavour profile analysis panel members were; David Rector, Bob Andrews, Marguerite Ferguson, Pat Melnychuk, Nola Low, Mary Tweedie, Dianne Sergy, Margaret Wanke, and Veronica Clough. We thank Leigh Noton of Alberta Environment for his advice and assistance and the staff of the Alberta Environment Millwoods Laboratory for sample collection. REFERENCES APHA-AWW A· WEF (1992). Standard Methods for the Examination of Water and Wastewater, 18th Edition, American Public Health Association. ASTM (1968). Manual on Sensory Testing Methods. American Society For Testing and Materials Special Technical Publication No. 434. Bartels, J. H. M., Brady, B. M. and Suffel, I. H. (1987). Training panelists for the flavor proflle analysis method. JA WWA, 79, 26• 32. Brownlee, B. G., Kenefick, S. L., Macinnis, G. A. and Hrudey, S. E. (1995). Characterization of odorous compounds from bleached kraft pulp miD effluent, Wat. Sci. Tech., 31(11) (this issue).
A chemical and sensory study of odour compounds
21
Brownlee, B. G., Macinnis, G. A. and Noton. L. R. (1993). Chlorinated Anisoles and Veratroles in a Canadian River Receiving Bleached Kraft Pulp Mill Eflluent - Identification, Distribution. and Olfactory Evaluation. Environmelllal Science and Technology. 27.2450-2455. Cook, W. H., Farmer, F. A.• Kristiansen, O. E.• Reid, K., Reid, J. and Rowbottom, R. (1973). The effect of pulp and paper mill eflluents on the taste and odour of the receiving water and the fish therein. Pulp and Paper Magazine of CanadJJ, 74, C97-106. Domtar Fine Papers Ltd. (1971). Effect of Pulp and Paper Mill Effluents on the Taste and Odour of Water and Fish. Pulp and Paper Pollution Abatement, Environment Canada. Project Report. Headley, J. V. (1987). GCIMS identification of organosulphur compounds in environmental samples. Biomedical and Environmental Mass Spectrometry, 14. 275-280. Kenefick, S. L., Brownlee, B.. Hrudey, E., Gammie. L. and Hrudey, S. E. (1994). Water Odour Athabasca River February and March, /993. Project Report No. 42. Northern River Basins Study, Edmonton, Alberta. Kenefick, S. L., Brownlee, B.• Hrudey. E., MacInnis, G. and Hrudey, S. E. (1995). Project 4413-Cl Water Taste and Odour Study (Post-AIPac). Submitted to the Northern River Basins Study Board, Edmonton, Alberta. Kovacs, T. G. and Voss, R. H. (1986). Factors influencing the effect of bleached kraft mill effluents on drinking water quality. Water Research, 20(9),1185-1191. Krasner, S. W. (1988). Aavor-profile analysis. An objective sensory technique for the identification and treatment of off-flavors in drinking water. Wat. Sci. Tech., 20(819). 31-36. NystrOm, A., Grimvall, A., KranlZ-Riilcker, C., Slivenhed, R. and Akerstrand, K. (1992). Drinking water off-flavour caused by 2,4,6-trichloroanisole. Wat. Sci. Tech.. 25(2). 241-249. Paasivirta, J., Knuutinen, J., Tarhanen, J., Kuokkanen. T.• Surma-Abo. K., Paukku, R., KlIliriliinen, H.. Lahtiperll, M. and Veijanen, A. (1983). Potential off-flavour compounds from chlorobleaching of pulp and chlorodisinfection of water. Wat. Sci. Tech.. 15,97-104. SUffer, I. H., Brady, B. M., Bartels, J. H. M.• Burlingame, G., Mallevialle, J. and Yohe, T. (1988). Development of the flavor profile analysis method into a standard method for sensory analysis of water. Wat. Sci. Tech.. 20(819), 1-9. Wigilius, B., Bor~n, H., Grimvall, A., Carlberg, G. E.. Hagen. I. and Brogger, A. (1988). Impact of bleached kraft mill eflluents on drinking water quality. Science ofthe Total Environment. 74, 75-96. Wong, A., Voss, R. H., Kovacs, T. G. and Dorica, J. G. (1985). Drinking water organoleptic quality as influenced by biologically treated bleached kraft mill eflluent, Journal ofPulp and Paper Science. 11(6), 161-166.
Pergamon
Welt. Sci T.ch. Vol. 31. No. 11. pp. 1S-21.
Copyright C
199~
199~.
IAWQ
Printed iD Great BritaiD. All rigbll reserved.
0273-1223(95)00450-5
0273-1223,o}~ S9'~0 +
0-00
A CHEMICAL AND SENSORY STUDY OF ODOUR COMPOUNDS IN THE ATHABASCA RIVER, ALBERTA, CANADA S. L. Kenefick*, B. G. Brownlee**, T. R. Perley* and S. E. Hrudey* • Environmental Health Program. University ofAlberta, Edmonton. AB. T6G 2G3. Canada •• National Water Research Institute. 867 Lakeshore Road, Burlington. ON. L7R 4A6 Canada
ABSTRACT As of early 1993. the Athabasca River received effiuent from one bleached kraft pulp mill (Mill A). three cbemithermomcebanical pulp and paper mills. one oil sands extraction and upgrading plant and a number of municipal effiuents. In the latter half of 1993 a second bleached kraft pulp mill (Mill B) began operation midway along the river. An investigation was carried out to cbaracterize the odours in the river water using both chemical and sensory methods. in a pre- and post-operational study of the second bleacbed kraft mill. Both surveys were earned out under ice during low flow conditions. Samples were analysed by gas cbromatograpby-mass spectrometry after extraction using a closed-loop stripping apparatus (CLSA). In addition, sensory analysis by flavour profile panel and olfactory gas chromatography of the CLSA extracts were performed. In 1993 all analytical methods conftrmed that compounds characteristic of bleached kraft mill effiuent were deteCtable for more than 950 km downstream from Mill A and that this effiuent was the major source of odour to the Athabasca River. Cbemical and sensory results for 1994 samples (collected after Mill B began operation) indicated a decrease in the impact of Mill A compared with the 1993 survey. The Mill B effiuent bad distinctive odour but its impact on the river was difficult to detect due to dilutioo and background odour from the Mill A effiuent.
KEYWORDS Bleached kraft pulp mill; chemithermomechanical pulp mill: closed-loop stripping; flavour proftle analysis; odour; olfactory gas chromatography
INfRODUcnON Bleached kraft pulp mill effluents are known to contain odorous compounds that have the potential to cause taste and odour problems for downstream users (Cook et aI., 1973; Kovacs and Voss, 1986; Paasivirta t't aI., 1910; Wigilius t't al., 1988; Wong et al., 1985). Such effluent has been reported to impair the taste and odour of drinking water at eft1uent concentrations ranging from 0.1 to 0.4% (Kovacs and Voss, 1986). Although most of the compound-specific work on pulp mill effluent odours has focused on chlorinated compounds, the literature does not provide a reliable basis for attributing odour problems primarily to 15
S. L. KENEFICK et al.
16
chlorinated compounds. Work by Domw Fine Papers (1971) to identify the most odorous process streams did not identify specific compounds but the process streams identified were rich in odorous organosulphur compounds. The production of these compounds is independent of bleaching practices and their control depends on good in-plant spill control and efficient wastewater treatment. Headley (1987) found odorous suiphones, sulphides and thiophenes in pulp mill effluents. He reported that sulphones and sulphides originated during biological treatment of kraft mill waste streams and thiophenes were formed as a result of the lignin digestion process, in which sodium sulphide and sodium hydroxide are used. Pulp mill effluent contributions to river water levels of phenols and trichlorophenol are also considered important after recent Swedish reports of bio-methylation of trichlorophenol to 2,4.6-trichloroanisole, an extremely potent source of musty odours in water and fish (Nystrom et ai., 1992). In early 1993, the Athabasca River received effluent from one bleached kraft pulp mill (Mill A), three chemithermomechanical pulp and paper mills, one oil sands extraction and upgrading plant and a number of municipal effluents. In the latter half of 1993 a second bleached kraft mill (Mill B) began operation midway along the river. A more complete description of the study area can be found in Brownlee et ai. (1995). Preliminary studies in 1991 had shown that odour compounds including 2,4,6-trichloroanisole and several chlorinated veratroles persisted for more than 1000 km downstream of Mill A (Brownlee et ai., 1993). Our investigation was carried out to further characterize the odours in the river water using both chemical and sensory methods. in a pre- and post-operational study of the second bleached kraft mill. The pre-operational survey was conducted in February and March of 1993 (Kenefick et ai., 1994) and the second or "post• operational" phase was conducted in February and March of 1994 (Kenefick et ai., 1995). This second phase of the study was also carried out after the condensate recovery system at Mill A was upgraded and its bleaching process was changed from 45% to 100% chlorine dioxide substitution. The primary objectives of this study were: I) Carry out comprehensive chemical, olfactory gas chromatographic. and flavour profile analyses to describe odours and their persistence in the Athabasca River; 2) Carry out this study before and after Mill B begins operation to determine the impact of the second mill; 3) Evaluate the usefulness of each analytical method (chemical, olfactory, flavour profile) for this kind of study and the overall value of using all three methods. METHODS The materials collected for both phases of this study included industrial effluent, municipal effluent, tributary, mainstream river and treated water samples. All water and effluent samples were collected at the time-of-travel of the Athabasca River as in Brownlee et ai. (1995). All were collected with no headspace, without addition of a preservative, and were cooled until delivery to the laboratory within 24 hours of collection. An 8 I water sample was collected at each river. tributary, and treated water sampling point and 2 I samples of each effluent were collected. E1ayour profile analysis The basics of the Flavour Profile Analysis method have been listed in a number of sources (ASTM, 1968; Bartels, et ai., 19H7; Krasner, 1988; Suffet, et ai., 1988) and these have been standardized in standard method #2170 (APHA-AWWA-WEF, 1992). The panel used in this study was run in accordance with this method. Closed-JoQp strippin~ and
~as
chromatompby-mass spectromet[y
One litre replicate subsamples of each river, tributary and drinking water sample were analysed, as received. All effluent samples were diluted 20:1 with odour-free water prior to extraction of 11 subsamples. A known
A chemical and sensory study of odour compounds
17
mass of I-chlorodecane was added to each sample, as a recovery standard, prior to CLSA. The samples were closed-loop stripped onto 1.5 mg carbon filters for 2 hours (water bath temperature at 30°C and filter temperature at 50°C) using a Brechbiihler AG closed-loop stripping apparatus. The filters were then extracted twice with 10 III and once with 5 III of carbon disulfide:acetone (9: 1). A known mass of biphenyl• d lO was added to the extract as an injection standard. A 2 III sample was analysed by GC-MS using injection in the splitless mode and 10 III of doubly distilled isooctane was added prior to subsequent analyses by OGC and GCIFIO. The GC-MS analysis was performed in the selected ion monitoring (SIM) mode on an HP 5890 GC with an HP 5970 mass-selective detector and an HP 59940 Chemstation® data system. In 1993 the GC-MS conditions were: 80°C to 180°C at 4°C/min and then lQoC/min to 280°C; 08-1301 column, 30 m x 0.25 mm ID, 0.25 11m film thickness; column head pressure 100 kPa; injector temperature 250°C; detector temperature 300°C. In 1994, with the added organo-sulphur target compounds, oven program and injection conditions were: 35°C to 90°C at 4°C/min and then 90°C to 280°C at lQoC/min; cool-on-column injector at 35°C and column head pressure 70 kPa. OIfactoQ'
~as
chromatoernphy
The CLSA extracts were also analysed by olfactory gas chromatography (OGC) using a Hewlett Packard 51190 GC where the conditions used in 1993 were: 80°C to 280°C at 4°C/min; 08-1301 column, 30m x O.32mm 10, 0.25 11m film thickness; column head pressure 70 kPa; injector temperature 250°C. In 1994, oven program and injection conditions were modified to: 35°C to 90°C at 4°C/min and then 90°C to 280°C at 10°C/min; with a cool-on-column injector temperature of 35°C and column head pressure of 70 kPa. For OGC, the column was raised through a heated transfer unit from the GC oven to a glass detection cone (olfactory detector outlet, SGE International). A 2 III sample was injected and after the solvent peak had eluted, the outlet was continually monitored over time. Elution time, intensity and odour descriptor were recorded. A six point intensity scale was used (1 = very weak, 6 = very strong). Control samples of known standard concentrations were used to help establish consistency in descriptors and intensity values. Method blanks and filter blanks were also analysed. RESULTS ANO DISCUSSION A summary of FPA, OGC and GC-MS results for a number of river sites are provided in Tables I and 2. In 1993 the FPA results demonstrated a strong increase in odour for Athabasca River samples downstream of Mill A when compared with the upstream sample. Panelists used descriptors similar to those used for the Mill A effluent and some panelists recognized these samples as reminiscent of pulp mill odour. There was little indication of any reduced intensity of odour until 1200 km downstream from the mill discharge. Other effluent samples (non-pulp mill) were only slightly more odorous than the tributaries and all were less odorous than the Mill A effluent. The flavour panel results indicate that the Mill A effluent was most likely the dominant source of odour in the mainstream river. These findings are consistent with the work of Kovacs and Voss (1986) wherein they found that biologically treated bleached kraft mill effluent could impair drinking water odour (as perceived by a flavour panel) at effluent dilutions in river water as high as 300 to 1000 fold. Flavour panel members also found that odour in the Fort McMurray (950 km downstream) water supply decreased substantially from raw to treated and descriptors that had been primarily associated with the Athabasca River samples downstream of Mill A were not used for the treated water. The oil sands upgrading plant effluent (downstream of Fort McMurray) was recognized as very distinctive and relatively strong, but with the possible exception of a sulphur descriptor it was difficult to recognize any impact of this effluent on the Athabasca River after dilution. The treated water at Fort Chipeweyan (1200 km downstream) had a very strong chlorine odour that likely masked any other subtle odours that may have been present.
S. L. KENEFICK et a/.
18
Table I. Distribution of odour in the Athabasca River (February and March. 1993) Distance Downstream from Mill A (km)
FPAAverage Intensity'" (most common descriptor)
OGCpeaks consistently detected.... (average intensity)
Target Compounds detected by GC-MS (Concentration in ngIL)
upstream
0.4 (odorless)
IBMP(1.5) 246TCA(1.5)
Mill A BKME
2.1 (sewage)
IPMP(2) sulphur(3) 246TCA(4.5)
20
1.4 (sewage)
IPMP(4) sulphur(2.5) 246TCA(3.5)
IPMP(3) 246TCA(l) 345TCY(4)
240
1.5 (pulp & paper)
IPMP(2) 246TCA(4) geosmin(2)
IPMP(l) 246TCA(3) geosmin(2) 345TCY(33)
400
1.8 (pulp & paper)
IPMP(2) 246TCA(4)
IPMP(2) 246TCA(4) geosmin(3) 345TCY(27)
550
1.2 (pulp & paper)
IPMP(3) 246TCA(3)
IPMP(I) 246TCA(6) geosmin(2) 345TCY(27)
650
1.6 (septic)
IPMP(2) 246TCA(2.5) geosmin(1.5)
246TCA(7) geosmin(2) 345TCV(24)
950
1.4 (pulp & paper)
IPMP(1.5) 246TCA(3)
246TCA(13) geosmin(2) 345TCY(26)
1100
1.2 (earthy)
IPMP(4) 246TCA(4) geosmin(2.5)
246TCA(5) geosmin(2) 345TCY(7)
1200
0.5 (earthy)
IPMP(2)
NID 246TCA(4) 345TCV(27)
NID
'wuh a scale maxImum of 3.0 ··Consistently detected = average intensity greater than I on a 6 point scale 246TCA = 2.4,6-trichloroanisole IPMP = 2-isopropyl-3-methoxy pyrazine IBMP = 2-isobutyl-3-methoxy pyrazine 345TCV = 3.4,5-trichloroveratrole Flavour profile result~ were similar in 1994 except that Mill A effluent odour was less intense and the odour descriptors changed to include woody and resinous. Mill B effluent had a similar odour but could not be differentiated in the river due to dilution and the background odour from Mill A effluent. The CLSNGC-MS analytical results demonstrated a number of interesting findings. The mainstream Athabasca River sample upstream of Mill A did not contain detectable levels of any of the target odour compounds other than geosmin (a biogenic compound) in 1993 or 1994. In the 1993 Athabasca River samples collected downstream of Mill A there was consistent detection of 3,4,S-trichloroveratrole (3,4,5• TCY) as far as 1100 km downstream. Geosmin and 2,4.6-trichloroanisole (2,4,6-TCA) were also detected in most mainstream samples downstream of Mill A (including Fort McMurray raw and finished waters). In 1994 the mainstream river samples collected downstream of Mill A did not show detectable levels of the target compounds that were found in the 1993 study. However, given the sulphurous odours detected using OGC in 1993. the 1994 GC·MS list of target compounds was expanded to include a number of thiophenes and thioethers (organosulphides). Low levels of thiophene and/or dimethyl disulphide (DMDS) were found as far as 750 km downstream of Mill A. This may have been caused by the presence of sewage treatment plant effluent in the Mill A discharge. The other three sewage treatment plant effluents all contained
A chemical and sensory study of odour compounds
19
thiophene or OMOS. However, the three biologically treated CTMP effluent samples also contained low levels of these sulphur compounds, indicating that a variety of biological treatment processes can result in their formation. CLSA analysis for target compounds did not provide evidence of causative agents which result in "pulp mill" odours. Table 2. Distribution of odour in the Athabasca River (February and March, 1994) Distance Downstream from Mill A (km)
FPA Average
Intensity· (most common descriptor)
OGe peaks consistently detected·· (average intensity) lPMP(2) IBMP(2)
Target COmpounds detected by GC-MS (Concentration in ngIL)
upstream
0.4 (odorless)
MillABKME
1.3 (septic)
NID
20
1.5 (septic)
sulphur(3) lPMP(3) IBMP(2) 246TCA(2)
240
1.3 (septic)
lPMP(3) IBMP(4)
thiophene(2) 01105(9)
400
1.3 (septic)
IBMP(2)
thiophene(1 ) geosmin(8)
550
1.3 (septic)
IPMP(2)
01105(2)
MillBBKME
1.4 (rubber/chemical)
650
1.4 (septic)
MIB(4)
750
1.5 (woody)
N/D
950
1.6 (septic)
IBMP(3) MIB(2) geosmin(2)
N/D
1100
1.0 (septic)
lPMP(2) IBMP(3)
N/D
1200
1.1 (septic)
N/D
N/D
sulphur(2) lPMP(2) MIB(4)
geosmin(9)
NID 01105(28)
NID O11OS(9) thiophene(3) 01105(12)
'WJih a scale maxImum of 3.0 ··deteeted = Intensity greater than 1 on a 6 point scale 246TCA = 2.4,6-trichloroanisole IPMP = 2-isopropyl-3-mcthoxy pyrazinc IBMP = 2-isobutyl-3-mcthoxy pyrazinc 34STCV 3.4,S-trichloroveralrolc MIB = 2-mcthylisobomcol DMDS = dimethyl disulphide
=
In the initial year of this study the predominant odours by oac in Mill A effluent were the 2.isopropyl-3• methoxy pyrazine (IPMP), sulphurous, and 2.4,6·TCA odours. One analyst also detected woody, sewage and spicy odours. The chemithermomechanical pulp mill effluents were generally low in odour but had some distinctive odour peaks described as woody, cardboard and sulphurous. The municipal wastewater treatment plant effluents also had low intensities but had a variety of distinctive odour peaks not found in the other samples, such as waxy, flowery, soapy, woolly, spicy and cucumber. In 1993 mainstream Athabasca samples (including the site upstream from Mill A), oac showed moderate to strong levels of 2.4,6-TCA and IPMP for 1100 km downstream of Mill A. Two unidentified musty cork odour peaks also showed up quite consistently at weak to moderate intensities. Geosmin occurred quite consistently at trace to moderate levels
S. L. KENEFICK et al.
20
and 3.4.5-TCV showed up occasionally at average intensities less than 1. Sulphurous and hydrocarbon odour peaks were also frequently noted. being strongest in the upper reaches and diminishing to non-detectable levels by Fort McMurray. The 1994 OGC results for the Mill A effluent differed sharply with those obtained for the same effluent in the 1993 study. The major process changes. including greater condensate recovery and 100% chlorine dioxide bleaching at Mill A, appear to have led to a significant reduction in odour peaks detectable with olfactory GC. The Mill B effluent sample showed weak levels of a sulphurous odour, and an IPMP odour and moderate levels of the camphorous odour of 2-methylisoborneol (MIB). In 1994 the mainstream Athabasca samples. including the site upstream from Mill A, showed weak to moderate levels of the musty pyrazine (lPMP or IBMP) odours as far as 1100 kIn downstream of Mill A. The musty. camphorous odour of MIB was detected in mainstream samples downstream of Mill B and as far downstream as Fort McMurray. The finished water sample collected at Fort McMurray had no detectable odour peaks and the Fort Chipewyan treated water had only a 2.4,6-TCA odour peak. CONCLUSIONS I) In 1993. and again in 1994, Kraft Mill A effluent was the major source of odour to the river. Low temperatures and ice cover allowed the odour to be transported long distances. This odour was detected at distances greater than 950 kIn downstream from the source. but the odour disappeared upon conventional water treatment, possibly due in part to masking by chlorine.
2) Mill B effluent did have a distinctive odour but its impact on the river was difficult to detect due to dilution and background odour from the Mill A effluent. CLSA analysis for target compounds did not provide evidence of causative agents for either effluent and a further expansion of the target compound list is required before conclusive. compound-specific tracing studies will be possible. 3) This paper summarizes olfactory GC, CLSNGC-MS and FPA results. but the three methods were difficult to correlate. The variety and intensities of odour peaks reported in OGC analyses gave a semi• quantitative statement about spatial distribution of odours. Flavour panel results gave an overall odour intensity for each sample and probably yielded the most useful information, but these FPA results could not be compared directly with chromatographic results. Unfortunately the current literature list of compounds responsible for the characteristic "pulp mill" odour was not applicable to the specific effluents being surveyed in this study. Unless the target compound list is redeveloped for each specific effluent, the three methods are difficult to link and they will remain difficult to correlate. ACKNOWLEDGEMENTS Funding for this research was provided by the Northern River Basins Study and the Natural Sciences and Engineering Council of Canada. The flavour profile analysis panel members were; David Rector, Bob Andrews, Marguerite Ferguson, Pat Melnychuk, Nola Low, Mary Tweedie, Dianne Sergy, Margaret Wanke, and Veronica Clough. We thank Leigh Noton of Alberta Environment for his advice and assistance and the staff of the Alberta Environment Millwoods Laboratory for sample collection. REFERENCES APHA-AWW A· WEF (1992). Standard Methods for the Examination of Water and Wastewater, 18th Edition, American Public Health Association. ASTM (1968). Manual on Sensory Testing Methods. American Society For Testing and Materials Special Technical Publication No. 434. Bartels, J. H. M., Brady, B. M. and Suffel, I. H. (1987). Training panelists for the flavor proflle analysis method. JA WWA, 79, 26• 32. Brownlee, B. G., Kenefick, S. L., Macinnis, G. A. and Hrudey, S. E. (1995). Characterization of odorous compounds from bleached kraft pulp miD effluent, Wat. Sci. Tech., 31(11) (this issue).
A chemical and sensory study of odour compounds
21
Brownlee, B. G., Macinnis, G. A. and Noton. L. R. (1993). Chlorinated Anisoles and Veratroles in a Canadian River Receiving Bleached Kraft Pulp Mill Eflluent - Identification, Distribution. and Olfactory Evaluation. Environmelllal Science and Technology. 27.2450-2455. Cook, W. H., Farmer, F. A.• Kristiansen, O. E.• Reid, K., Reid, J. and Rowbottom, R. (1973). The effect of pulp and paper mill eflluents on the taste and odour of the receiving water and the fish therein. Pulp and Paper Magazine of CanadJJ, 74, C97-106. Domtar Fine Papers Ltd. (1971). Effect of Pulp and Paper Mill Effluents on the Taste and Odour of Water and Fish. Pulp and Paper Pollution Abatement, Environment Canada. Project Report. Headley, J. V. (1987). GCIMS identification of organosulphur compounds in environmental samples. Biomedical and Environmental Mass Spectrometry, 14. 275-280. Kenefick, S. L., Brownlee, B.. Hrudey, E., Gammie. L. and Hrudey, S. E. (1994). Water Odour Athabasca River February and March, /993. Project Report No. 42. Northern River Basins Study, Edmonton, Alberta. Kenefick, S. L., Brownlee, B.• Hrudey. E., MacInnis, G. and Hrudey, S. E. (1995). Project 4413-Cl Water Taste and Odour Study (Post-AIPac). Submitted to the Northern River Basins Study Board, Edmonton, Alberta. Kovacs, T. G. and Voss, R. H. (1986). Factors influencing the effect of bleached kraft mill effluents on drinking water quality. Water Research, 20(9),1185-1191. Krasner, S. W. (1988). Aavor-profile analysis. An objective sensory technique for the identification and treatment of off-flavors in drinking water. Wat. Sci. Tech., 20(819). 31-36. NystrOm, A., Grimvall, A., KranlZ-Riilcker, C., Slivenhed, R. and Akerstrand, K. (1992). Drinking water off-flavour caused by 2,4,6-trichloroanisole. Wat. Sci. Tech.. 25(2). 241-249. Paasivirta, J., Knuutinen, J., Tarhanen, J., Kuokkanen. T.• Surma-Abo. K., Paukku, R., KlIliriliinen, H.. Lahtiperll, M. and Veijanen, A. (1983). Potential off-flavour compounds from chlorobleaching of pulp and chlorodisinfection of water. Wat. Sci. Tech.. 15,97-104. SUffer, I. H., Brady, B. M., Bartels, J. H. M.• Burlingame, G., Mallevialle, J. and Yohe, T. (1988). Development of the flavor profile analysis method into a standard method for sensory analysis of water. Wat. Sci. Tech.. 20(819), 1-9. Wigilius, B., Bor~n, H., Grimvall, A., Carlberg, G. E.. Hagen. I. and Brogger, A. (1988). Impact of bleached kraft mill eflluents on drinking water quality. Science ofthe Total Environment. 74, 75-96. Wong, A., Voss, R. H., Kovacs, T. G. and Dorica, J. G. (1985). Drinking water organoleptic quality as influenced by biologically treated bleached kraft mill eflluent, Journal ofPulp and Paper Science. 11(6), 161-166.