4 Effects Characterization
The previous review of the ecological risks associated with chlorine dioxide bleaching (Solomon et al., 1993) concluded that substitution of chlorine dioxide for chlorine results in a reduction in toxicity of pulp mill effluents and a reduction in adverse ecological effects. The report stated that "mills bleaching with high chlorine dioxide substitution (100%), employing secondary treatment and with receiving water dilutions typical of most mills in North America, present an insignificant risk to the environment from organochlorine compounds." The previous review also identified research needs related to effects including:
This chapter begins with a brief summary of previous knowledge of the effects of pulp mill effluents (BKME) on aquatic biota. It describes recent research on the causes of these effects and on the relationship of effects to pulping, bleaching and effluent treatment technologies, particularly ECF vs TCF bleaching. New data were not found that addressed directly the questions of the release of sediment-borne compounds, the effects of chlorate in freshwaters, or the nature of effects on severely affected species.
Land-mark field studies in Sweden demonstrated that the discharge of BKME can cause important detrimental effects on fish populations (see review by Hodson et al., 1996a). The most obvious effects were caused by older mills, characterized by bleaching with elemental chlorine, poor spill and carbon recovery systems, and little or no effluent treatment. In fact, one of the mills in the study was operating very poorly because it was being up-graded during the study. Fish downstream accumulated chlorinated organic compounds and showed physiological responses indicative of stress and of changes in rates of metabolism of xenobiotic substances, increased excretion rates of compounds found in BKME, increased prevalence of gross deformities, reduced recruitment, and shifts in fish population and community structure. Improvements to mill processes, replacement of chlorine with ClO2 as a bleaching agent, and secondary effluent treatment sharply reduced the discharge of chlorinated compounds and the contamination of fish. Many of the signs of toxicity, such as gross deformities, were eliminated, but some physiological responses remained, particularly increased rates of metabolism of xenobiotics. The Swedish EPA initially interpreted the reduction in effects as a response to reduced loadings of chlorinated organic compounds. However there were no chemical analyses of non-chlorinated compounds in effluent or fish, (Tana and Lehtinen 1996; Owens 1991), so their contribution to toxicity was unknown.
The effects on fish of BKME were later confirmed in North America, but the newer studies led to a different interpretation -- effects were likely due to natural constituents of wood. The North American studies also demonstrated that fish downstream of some bleached kraft mills showed differences in the serum concentrations of reproductive steroid hormones. Coupled with changes in fish population demographics, such as reduced gonad size, reduced fecundity, delayed maturity, and increased size at maturity, the hormone responses were consistent with the idea that population impacts of BKME are due to impaired reproduction.
The North American studies differed from the Swedish studies in that effects on fish were not unique to kraft mills nor to chlorine bleaching. Effects on steroid hormone concentrations and induction of MFO enzymes were also observed in fish downstream of kraft, sulfite, thermomechanical pulping (TMP) and chemi-thermomechanical (CTMP) mills not using bleaching. Increased rates of deformities and skin lesions seen in Sweden were less obvious in North America (Couillard et al., 1995). Further, many of the physiological responses attributed to effluents could be accounted for as responses to seasonal variations in fish physiology due to natural gradients of habitat and growth rates (Hodson et al., 1992a). Physiological measurements are also biased by capture stress, which cannot be alleviated by storing fish in cages before sampling (McMaster et al., 1994; Jardine et al., 1996a; 1996b). Effects on demographics were sometimes confused by the variable growth response; reproduction enhanced by nutrient-stimulated growth rates may have obscured impairment of reproduction by toxicity (Gagnon et al., 1995; 1996).
The importance of secondary treatment was also uncertain. At several mills in Ontario where secondary treatment and increased ClO2 substitution were initiated, effects on fish in receiving waters were reduced but still evident (Munkittrick et al., 1992). More importantly, studies of some bleached kraft mills in Western Canada found few effects (Swanson et al., 1994; Kloepper-Sams et al., 1994), indicating that effluent toxicity was not universal. The technological basis for differences in effects are not at all clear, but there are an increasing number of studies demonstrating the benefits of both ClO2 substitution and enhanced effluent treatment, in terms of reduced chemical contamination and reduced physiological responses of fish (e.g., Oikari and Holmbom, 1996).
Field studies can demonstrate which technologies are not uniquely responsible for observed effects on fish; bleaching with chlorine or chlorine dioxide is not a necessary process for pulp mill effluent to have effects on fish. The nature of field studies, in which different observations can only be correlated, and in which numerous factors act simultaneously and independently, prevents the development of clear cause-effect relationships. Early conclusions ascribing BKME effects solely to organochlorine compounds were incorrect.
This section reviews recent advances in the understanding of the mechanisms, causes and consequences of two frequently-observed responses of fish to pulp mill effluents: changes in metabolism by enzymes and impaired regulation of reproductive steroid hormones. Knowledge of mechanisms provides a basis for assessing whether chlorinated compounds play a role in these responses, and whether ECF bleaching is an important source.
Mixed-function oxygenases (MFO) are enzymes that oxygenate endogenous and exogenous low molecular weight substrates. Activity of MFO enzymes, such as is measured by ethoxyresorufin-O-deethylase (EROD) activity, increases following exposure of fish to pulp mill effluents. This increase in activity is called induction. Significant MFO induction indicates exposure and potential adverse effects; conversely, the absence of induction is evidence that either effluent does not contain sufficient inducers to elevate MFO activity, or that MFO inhibition has occurred due to high concentrations of other chemicals (e.g., resin acids) (Hodson et al., 1996c).
The potential adverse effects associated with MFO induction may be due to unrelated toxicities of inducing compounds (i.e., induction simply indicates the degree of exposure) or to the oxygenation of compounds by MFO enzymes. Oxygenation of aromatic compounds creates reactive intermediates such as epoxides and alcohols that may be conjugated to amino acids. These conjugates are more polar than the parent compound, and are readily excreted in the bile. Hence, increased activity can reduce toxicity by reducing exposure. Production of reactive intermediates can also be harmful. Some cause strand breaks in DNA or react readily with nucleic acids, forming "adducts" that can cause mutations, impaired replication of DNA, and sometimes cancer. Adducts have recently been reported for the first time in chinook salmon (Oncorhynchus tshawytscha) exposed naturally or in the laboratory to BKME (ECF, 8-d aerated stabilization pond) from a mill at Prince George, B.C. (Wilson, 1997). An elevated frequency of micronuclei in erythrocytes of chinook salmon exposed for 30 d to effluent from the same mill (Easton et al., 1997) is also consistent with a mutagenic response. The effect was observed at 4% (v/v) effluent or greater. Another consequence of reactive intermediates is the oxidation of other substrates, such as polyunsaturated lipids. Where exposures are low, the anti-oxidant defense system may neutralize the reactive intermediates; this system includes enzymes that catalyze reducing reactions, reductants such as glutathione, and antioxidant vitamins, such as vitamin C and E. If there is sufficient oxidation, a condition of oxidative stress may occur, in which lipids begin auto-peroxidation (rancidity), leading to cell membrane damage and cell death.
Prolonged MFO induction by TCDD, and hence prolonged activation of the cypIA1 gene has been associated with a suite of toxic effects known as receptor mediated toxicity (RMT), which includes involution of the thymus, wasting, and cell membrane damage (Okey et al., 1994). In fish the effects are most evident in larvae, which exhibit "blue-sac" disease, characterized by edema (particularly of the yolk sac), cranio-facial deformities, fin erosion, and hemorrhaging, particularly around the heart. The condition is fatal (Spitsbergen et al., 1991). RMT can be caused by other persistent inducers such as co-planar PCBs, chlorinated dioxins, and chlorinated furans, and potency for "blue sac" disease is similar to potency for MFO induction (Walker and Peterson 1991; Parrott et al., 1995a). Prolonged exposure to retene (methyl-isopropylphenanthrene) also causes prolonged MFO induction and symptoms of blue sac disease in developing trout (N. Fragoso, S. Billiard, and P. Hodson, unpublished data).
Despite an association
of MFO induction with different mechanisms of toxicity, the significance
of induction in wild fish is difficult to interpret. The duration of chemical
exposure and MFO induction is usually unknown and the extent of induction
associated with toxicity has not been characterized. Further, toxicity
is usually assessed with laboratory tests of juvenile stages of trout,
while induction in field studies is measured in adults of other species.
For example, high concentrations of dioxins were found in whitefish near
a mill in Saskatchewan, without apparent adverse effects (Kloepper-Sams
and Benton 1994).
The disruption of reproductive steroid hormone regulation in fish (Munkittrick et al., 1991) provides a plausible mechanism for observed reproductive impairment, as shown by delays in age to maturity, smaller gonad size, and reduced fecundity. Research since 1993 has indicated that one of the candidate compounds for causing these effects is ß-sitosterol, a non-chlorinated natural constituent of wood, and the most abundant plant sterol released during pulping (Stuttaford, 1995). Sitosterol can be metabolized by soil bacteria to a variety of androgenic and estrogenic compounds, and this process has been shown to cause changes in secondary sexual characteristics of tropical fish species (e.g., Heterandria sp., Davis and Bartone, 1992). Debarking effluent, abietic acid, and sitosterol have been found to cause vitellogenesis in fish, a sign of an estrogenic response (Mellanen et al., 1996), and injected sitosterol interferes with reproductive steroid hormone regulation of goldfish (Carassius auratus; MacLatchy and Van Der Kraak, 1995). While injection of goldfish with sitosterol causes a change in serum concentrations of reproductive steroid hormones, it is not clear how sitosterol is acting and whether concentrations in final effluent are sufficient to cause the observed reproductive impacts on fish. Very little sitosterol seems to survive past secondary treatment (<10 µg/L; Stuttaford, 1995; also see section 3.4.2). Further, while endocrine disruption by pulp mill effluent exposure was found consistently at some mills, the point in the synthetic pathway of reproductive steroid hormones where pulp mill effluent exposure causes impairment is still not well defined (McMaster et al., 1995).
A review by McMaster et al. (1996a) of a series of field studies indicated that reproductive impairment, as shown by changes in hormone concentrations or rates of hormone synthesis, is frequently found in fish downstream of pulp mills. However, the pattern of response is not universal, because fish exposed to BKME in the Androscoggin River in Maine showed elevated serum testosterone concentrations, in contrast to fish sampled at other BKME-contaminated sites (McMaster et al., 1996a). Some species such as brown bullhead (Ictalurus nebulosus) were also less sensitive compared to white sucker and steroid effects could also be found in fish exposed to polynuclear aromatic hydrocarbons (PAH) from steel mills. Understanding the relationship of these responses to pulp mill technology will require an understanding of the biochemical basis of the effect and of the compounds causing those effects. However, it was obvious that reproductive effects could occur in the absence of ECF bleaching, as shown by reproductive responses of fish from the Winnipeg River, sampled downstream of a kraft mill not using chlorine compounds for bleaching.
No links have yet been demonstrated between MFO induction and impairment of reproductive responses (e.g., Swanson et al., 1994); they may have entirely separate causes and consequences. While some compounds causing induction also cause reproductive impairment (e.g., co-planar PCBs, TCDD, PAH; McMaster et al., 1996a), the mechanisms may be independent. In fact, increased estrogen in female fish during the final stages of sexual maturation "down-regulates" the cyp450 gene so that MFO induction is reduced.
Zacharewski et al. (1996) described a transfected yeast bioassay that reacts to estrogen-like compounds; this "reporter-gene" bioassay was developed as a screening tool. They tested undiluted black liquor from a bleached kraft pulp mill as well as a specific fraction isolated from the secondary-treated effluent from the same mill. In previous tests, both samples caused MFO induction in fish and, in this yeast bioassay, both samples demonstrated estrogen-like activity. However, it is unknown at this time whether the two responses involve the same chemicals.
The relationships between MFO induction and effects on fish of pulp mill effluent are as yet undefined. MFO induction and alterations in serum concentrations of reproductive steroid hormones appear to be separate and unrelated responses of fish to pulp mill effluents. While the hormones response may be the mechanism by which pulp mill effluents affect fish reproduction, the specific details are not completely worked out.The toxicilogical signifiance of MFO induction is related to prolonged elevation of MFO activity and prolonged gene activation. Both steroid effects and prolonged MFO induction can be caused by non-chlorinated natural constituents of wood produced by pulping, not by ECF bleaching.
There have been numerous studies demonstrating MFO induction in fish downstream of pulp mills, and exposed to effluents, diets, sludges, and sediments in the laboratory (Table 4-1). The consistency of the response and its possible connection to toxicity (see above) has prompted a search for inducing compounds and for pulp mill processes controlling their concentrations in effluent. The search first focused on chlorinated dioxins and furans because of their high potency for induction (Parrott et al., 1994), their occurrence in effluent from mills using chlorine bleaching (Berry et al., 1991), and the close relationship between dioxin equivalents in fish tissues and measures of MFO induction (Hodson et al., 1992b; Kloepper-Sams and Benton 1994). At mills where elemental chlorine was replaced by ClO2, dioxin and furan emissions were substantially eliminated, concentrations in sediments and biota decreased dramatically, and fisheries were reopened (see Sections 2.3.2; and 3.5). However, MFO induction in fish from receiving waters did not disappear at all locations (Munkittrick et al., 1992; van den Heuvel et al., 1996), and induction was found in fish exposed to effluents from all types of mills, including those not using chlorine or ECF bleaching (Munkittrick et al., 1994; Gagné and Blaise, 1993; Martel et al., 1994). These results were confirmed in a more controlled experiment in which bass (Micropterus salmoides) and catfish (Ictalurus punctatus) were exposed to BKME in experimental streams for periods of up to 263 d (Bankey et al., 1995). Fish exposed to 4-8% (v/v) effluent from a mill using 70% CIO2 substitution were induced 55-fold over controls, but induction was lost rapidly when the fish were transferred to fresh water. TCDD concentrations in effluent ranged from <6 - 20 pg/L, while TCDF concentrations were about four times greater. A subsequent exposure of fish to 4-12% (v/v) effluent after the mill switched to oxygen delignification and 100% ClO2 substitution also demonstrated MFO induction, but in this case, the potency was much less and induction was only 3-fold over controls. For the ECF effluent, TCDD concentrations were lower but exceeded the detection limit of 5 pg/L in 5% of samples; TCDF was detectable in 47% of samples. Hence, while dioxins and furans may have contributed to early measurements of induction, they were not the only cause, and they are certainly not the current cause.
Hewitt et al. (1996) concluded that resin acids, chlorophenols, aliphatic alkanes, plant sterols, chlorinated dimethyl sulfones and terpenes were not the cause of MFO induction in trout exposed to fractions of BKME; these compounds were found at high concentrations in effluent fractions that did not cause induction. Fish and cell culture bioassays of MFO induction by compounds tested singly confirm that chlorophenols, resin acids, and sitosterol are not inducers (Lehtinen, 1996; Hodson and Parrott, unpublished data), but defoamers cause slight induction at very high concentrations (Hodson and Parrott, unpublished data).
Pulping liquors, both weak black liquor from kraft pulping and spent liquor from alcohol pulping, contain high concentrations of MFO inducers, as shown by bioassay (Hodson et al., 1997). Small losses of black liquor to effluent could explain the induction potency of final effluent. Bleaching effluents may also be a large source of inducers, but secondary treatment systems can be an efficient means of reducing the MFO induction potential of effluents (Schnell et al., 1993; Martel and Kovacs, 1996a).
Inducers can be extracted from wood chips using dichloromethane, with about the same efficiency of removal as with alcohol pulping (Hodson et al., 1997), indicating that inducers are natural substances extracted from wood. The results are consistent with earlier studies of MFO induction in rainbow trout liver cells exposed to extracts of bleached and unbleached kraft and unbleached sulphite mill effluents, and in perch downstream of bleached and unbleached mills (Hanninen et al., 1991; Pesonen and Andersson 1992).
a ASB - Aerated Stabilization Basin; AS - Activated Sludge
Burnison et al. (1996) isolated MFO-inducing fractions of secondary-treated effluent from a mill using 50% ClO2 substitution. Bioassays with rainbow trout demonstrated that inducers were present in the aqueous phase and associated with dissolved organic matter and particulates (the fractions containing the largest amount). Inducers could be extracted from particulates by, methanol and from the aqueous phase by adsorption to C-18 resins. Extracts cleaned up on C-18 resin were separated by reverse-phase HPLC using methanol, and induction was isolated primarily to one fraction, along with standard marker substances with log KOWs of 4.6 - 5.1. Therefore the inducer appeared to be moderately hydrophobic, slightly soluble in water, and easily accumulated from water by fish. Compounds identified by preliminary scans with GC/MS included both PAHs (alkyl-substituted phenanthrene) and chlorine-substituted diarylethanes and diaryl ethenes (stilbenes) (Hodson, 1996). More recently, Burnison et al. (1997) reported that tri-chlorinated stilbenes were the most common constituent in HPLC fractions tested by MFO bioassays with cell cultures, although the identity of this compound must still be confirmed. In this case, 67 fractions were separated by HPLC, and activity was isolated to one fraction only. Engwall et al. (1997) have shown that polyaromatic substances extracted from bottom sediments in pulp mill effluent receiving waters contain high concentrations of MFO inducers that were chemically different from the dioxins and furans and the 15 priority PAH pollutants.
The terpenes, juvabione and dehydrojuvabione have recently been identified as inducers that were isolated from balsam fir (Martel et al., 1996c). In effluents from mills pulping balsam fir, the extent of induction was accounted for entirely by the concentrations of juvabione present. Juvabione is not found in other wood species commonly pulped in Canada, but it illustrates that a variety of natural wood constituents may be inducers. Other terpenes have been reported to not induce MFO activity (Hewitt et al., 1996).
Retene is an inducing compound found in pulp mill effluents, sludges from pulp mill treatment ponds, and in sediment downstream of pulp mills (Judd et al., 1995); its origin is thought to be microbial biotransformation of abietic and dehydroabietic acid under anaerobic conditions (see Sections 2.4 and 3.3). Bioassays have demonstrated that retene and other alkyl-substituted phenanthrenes cause 10 to 100-fold MFO induction in rainbow trout (Parrott et al., 1995b). Induction can be sustained during prolonged exposure to retene, but declines rapidly when fish are transferred to clean water (Fragoso et al., 1996), 1ikely due to retene metabolism and excretion; it is measurable in high concentrations in the bile of fish following exposure (Stuthridge et al., 1995). In fact, all of the wood extractives associated with biological responses have structures that appear to be susceptible to metabolic attack, conjugation and excretion.
MFO induction in fish is a common response to many pulp mill effluents, and not a characteristic unique to bleaching. MFO-inducing compounds in biologically-treated effluent are natural extractives of wood, their metabolic by-products, or their chlorinated derivatives. These inducing compounds appear to be non-persistent and easily metabolized and excreted by fish. The relative amounts of these compounds released by ECF or TCF effluents are unknown.
Numerous comparisons of effects on fish have been made among different pulping, bleaching and effluent treatment technologies. However, as was the case with Project Environment Cellulose I and II, studies conducted since 1993 have not clearly related specific pulp mill technology to specific effects.
Effluents from Ontario mills were tested for effects using 7-d Ceriodaphnia reproduction and survival tests and 7-d fathead minnow (Pimephales promelas) survival and growth tests (Robinson et al., 1994; Munkittrick et al., 1994; Servos et al., 1994; Van den Heuvel et al., 1994). Simultaneously, serum concentrations of reproductive steroid hormones and liver MFO activity were measured in male and female white sucker sampled downstream of the same mills and at reference sites. The results demonstrated that effluents from all types of mills cause potential reproductive effects and MFO induction in adult fish in receiving waters, but that effluents receiving secondary treatment were less toxic according to laboratory toxicity tests. There was insufficient information to really compare effects to pulp mill technologies, although some mills that caused effects (sulfite/mechanical mills) did not use any bleaching. Hence, the effects measured could not be associated with the presence or absence of chlorine. Of the four kraft mills having only primary treatment, one used a traditional bleaching sequence (CEH), and the others ClO2 substitution ranging from 16-30% to 100% (i.e., ECF). Effluent bioassays demonstrated that the ECF mill caused no effect on fathead minnow survival or growth and no effect on Ceriodaphnia survival and reproduction (Robinson et al., 1994). Mills using the traditional sequence and 50-60% substitution caused increased mortality rates of fathead minnow larvae; only effluent from the mill with 16-30% substitution inhibited Ceriodaphnia reproduction. For these 4 mills, there was no relationship between bleaching sequence and MFO induction, serum reproductive steroid hormone concentrations, and dioxin equivalents in white suckers from receiving waters. Differences in water use, receiving water dilution and process efficiencies prevented further useful analyses.
Reductions in effects on fish have been observed at some BKMs, but the relationship to changes in technology are not obvious. McMaster et al., (1996b) reviewed a series of studies of a mill at Jackfish Bay, Ontario, where reproductive steroid hormones were measured in serum of native white sucker exposed in situ, and where goldfish were exposed experimentally. Rates of hormone synthesis by isolated gonadal tissue of goldfish were reduced in early studies, but between May, 1993 and May, 1994, the extent of effluent effects on goldffsh hormones disappeared, coincident with unspecified changes to mill processes. While these included an increase in ClO2 substitution from 50 to 70%, the majority of changes were not made public, and effects on white sucker steroid hormones were still detectable in late 1993.
Field studies of pulp mill effluent effects on fish suggest responses of fish to effluent can be found with or without bleaching. Secondary treatment alleviated acute toxicity to fish and invertebrates in laboratory tests, but responses in some fish from receiving waters are still evident, although reduced. These studies demonstrate no other obvious relationships between pulp mill technologies and biological responses.
Tana et al., (1994) studied brackish water mesocosms treated with four effluents from mills with different bleaching sequences: CEHDED; O[C85D15][EO]DED; 2 mixed streams: (O[C60D40][EO]D) and (O[C60D40][EO]DED); and OPD. The acute lethalities of the 4 effluents were 23, 55, 27 and 100% v/v, respectively, with toxicity emission factors of 530, 70, 130 and 30, respectively. Mesocosms were exposed to effluent diluted either 200-fold or 1000-fold, and rainbow trout in the mesocosms were sampled after 8 weeks. Relative to controls, there was very little difference in MFO activity, with about a 50% induction observable only at the high concentration of the effluent from the mill with traditional bleaching. Lehtinen (1996), in a review of biochemical responses of fish in mesocosms, found that many effluents actually inhibited MFO activity, as well as activity of conjugating enzymes, at concentrations as low 0.25%, likely due to the presence of fatty and resin acids and chlorophenols. The inhibition of enzyme responses raises the possibility that comparisons of responses to technology may be confounded by "false negatives" due to toxicity of effluent components. Despite a large number of studies involving effluents from mills using many different bleaching sequences, there was no strong evidence that bleaching was the source of toxic components in final effluent (Lehtinen, 1996). The MFO responses of rainbow trout (measured as EROD activity) exposed to different whole mill effluents tested in mesocosms are compiled in Figure 16. Despite the fact that effluents from modern ECF and TCF processes generally show low or no acute toxicity, the EROD response tends to be lower than that of unexposed fish implying possible enzyme inhibition and sublethal toxicity. It is also noteworthy that peat bog water and urban sewage water show the same tendency.
Effluents from mills Kovacs and McGraw (1996) provided a detailed review of laboratory studies of effluent toxicity for studies published between 1991 and 1994. For both acute and chronic toxicity, there was evidence of an improvement in effluent quality, as shown by reduced effects on algae, invertebrates and fish, as mills upgraded pulping, delignification, bleaching, and waste treatment processes. However, the results did not allow clear conclusions about the benefits of specific technologies due to the numbers of changes made simultaneously, and the use of different test species and protocols by different authors. While laboratory tests can contribute to understanding the impacts by providing data from controlled conditions, data can often not be extrapolated to real situations because the tests are over-simplified.
Detailed mill sampling and laboratory bioassays demonstrated that the most potent source of MFO-inducing compounds within a bleached kraft mill, and a large source by volume, was the spent cooking liquor from kraft pulping (Schnell et al., 1993). Hence, bleaching is not essential for MFO induction, and inducers appear to be natural products derived from wood (see section 5.1.2). Nevertheless, the bleach plant (50% substitution) and first extraction stage effluents were much less potent MFO inducers than black liquor, but larger sources of inducers by volume. Possible explanations are carry-over of spent pulping liquor to the bleaching stage; a continued release of wood extractives or breakdown products of lignin by bleaching; or enhanced hydrophobicity and increased accumulation of extractives due to chlorination.
Schnell et al. (1993) also showed that the potency for MFO induction in fish of a variety of waste streams could be reduced significantly through enhanced biological treatment in lab-scale and pilot-scale treatment systems, specifically by increased solids and hydraulic retention times. However, treatment may simply transfer inducers to sludge. Sludges from the treatment system and downstream sediments of a bleached sulfite mill cause MFO induction in trout in sediment bioassays (Parrott and Hodson, unpublished data), and the highest concentrations of inducers in bleached kraft mill effluents were associated with particulates and colloidal material (Burnison et al., 1996). Lehtinen et al., (1991) found that sludges incorporated into fish food caused MFO induction in exposed fish, but induction was less when the sludge was pre-extracted with hexane.
Effluent from a variety of kraft mills, including those with chlorine bleaching, those with 50-100% substitution, and those with no bleaching at ale caused MFO induction in trout (Williams et al., 1996). Variation in induction among mills was least when differences in water use (effluent dilution) were removed from the comparison by normalizing to DOC. Williams found low coefficients of variation for comparisons of MFO induction potency and concentrations of AOX, chlorophenols, and resin and fatty acids. Coefficients of variation were best for DOC.
Effluent from five pulp mills on the Athabasca River system were surveyed for release of MFO inducing compounds using semi-permeable membrane devices (SPMDs) to collect samples (Parrott et al., 1996). SPMDs are 1 m x 2 cm flat tubes of polyethylene that contain 1.0 g of triolein, a fish lipid into which hydrophobic compounds can partition. In laboratory and field studies of two mills in Ontario, SPMDs accumulated significant amounts of MFO-inducing compounds from secondary-treated effluent from BKMs with 50% substitution, as shown by bioassays with mammalian cells in culture (Parrott et al., 1994). Induction by extracts of SPMDs paralleled induction in native fish from receiving waters, lab fish caged in receiving waters, and trout exposed to the same effluent in the laboratory. However, in the Athabasca River, none of the SPMDs installed in effluent discharge ponds or pipes, or in the river downstream of each mill, accumulated inducing compounds. In contrast, very high concentrations were accumulated from a "positive control", the effluent of a tar sands extraction facility.
The Athabasca mills included kraft, TMP and CTMP mills, with and without bleaching. Three mills were ECF and two mills were TCF, using alkaline peroxide and hydrogen peroxide as bleaching agents; all had secondary treatment. The absence of inducing compounds shows that pulp and paper mills employing either ECF or TCF bleaching sequences and secondary waste treatment are capable of producing final effluent without MFO inducing compounds. Because only final mill effluent was sampled, the technology responsible for the absence of inducing compounds could not be discerned.
Martel and Kovacs (1996b) tested the effects of different pulp mill technologies on the induction of MFO activity in trout fish by bioassays of effluents from a variety of pulp mills: kraft (N=7), CTMP (N=3), and TMP (N=3). For bleaching, four kraft mills used oxygen delignification and 100% C1O2 substitution, two had similar systems but did not use delignification, and the last had two pulping lines using partial substitution (OD60C40EoHD and CSOD50EHD). The CTMP mills brightened the pulp with hydrogen peroxide, while the TMP mills used hydrosulphite. Over a concentration range of 1-10%, primary-treated effluent caused MFO induction, but the extent of induction was not correlated to the type of pulping, wood furnish, pulp bleaching or brightening; there were no systematic differences in response to effluents between mills that did not use ClO2 and those that did. Similarly, secondary treatment by aerated lagoons or activated sludge reduced induction relative to untreated effluent, but the extent of reduction was unrelated to the technology of waste treatment. While it is clear that effluents from mills using ECF bleaching caused no greater induction than effluents from mills using no bleaching at all, the relationship of induction to technology remains enigmatic. One reason is that comparisons were based only on the extent of MFO induction in fish exposed to a limited range of test concentrations, some of which approached lethal levels.
A series of fathead minnow life-cycle bioassays and 7-d Ceriodaphnia tests conducted by PAPRICAN and NCASI (summarized in Table 4-2) demonstrated the benefits of process improvements on the chronic toxicity of BKME and the relative toxicity of TMP effluent. Kovacs et al., (1995b) measured the chronic toxicity to fathead minnows of a secondary-treated effluent from a bleached kraft mill pulping mixed softwoods (predominantly lodgepole pine) with modified continuous cooking and bleaching with D/C Eo [DED] after oxygen delignification; ClO2 substitution was 45%. Dioxins in the effluent were below detection (3 pg/L for 2,3,7,8-TCDD) but tetrachloro, hexachloro, heptachloro and octachloro furans were measurable, with mean concentrations of 2,3,7,8-TCDF of 19 pg/L (ppq). There were no effects on survival and growth of minnows, but reproduction was impaired: there was a delay in sexual maturity and a reduction in the numbers of eggs produced per female. The authors estimated that effects occurred at concentrations of 2.5% (v/v) or greater. Ceriodaphnia reproduction was also affected, but at much higher concentrations. The reproductive effects were very similar to those observed in wild fish downstream of kraft mills, with and without secondary treatment, and with and without ClO2 substitution (Gagnon et al., 1995; Munkittrick et al., 1994; Adams et al., 1992).
Reproduction of fish returned to normal when effluent from the same plant was tested after a variety of mill improvements (Kovacs et al., 1996); effluent concentrations impairing Ceriodaphnia reproduction increased from 36% to 68% (v/v). Changes included a change to 100% ClO2 substitution (now D1OOEop[DED]), improved brown stock washing, addition of anthroquinone, improved recovery and treatment of condensates, improved steam stripping of volatiles, reduced black liquor losses, a 35% reduction in defoamer use, and a change in type of pitch dispersants. The effluent was treated with primary clarification of all streams except the bleach plant effluents. Where they had previously been treated in a clarifier, they now were combined with primary-treated effluent for secondary treatment in a 5-d retention aerated lagoon. Secondary treatment included the domestic sewage from a town of 9,000, so that ammonia was no longer added to promote treatment. The lagoons were dredged and aeration reduced to increase the quiescent zone and reduce total suspended solids. The final effluent (70,000 m3/d) was mixed with 35,000 m3/d of cooling water before discharge. Previously, about 15,000 m3/d of cooling waters were passed through treatment, while 20,000 m3/d was by-passed. Now, all cooling waters by-pass treatment and are mixed with final effluent, which now contains a greater proportion of clean cooling water. The concentrations of color, COD, BOD, TSS, and AOX, declined by 30-80%, with the largest change in AOX; no measurements of dioxins and furans were reported. While effluent quality certainly improved, the role of changes in bleaching technology in toxicity reduction could not be assessed because there were too many simultaneous changes in mill operations and effluent treatment. As well, the contribution of urban sewage to toxicity could not be discriminated with these data.
a Bleaching sequence codes: D-chlorine dioxide; O-oxygen delignification; H-hypochlorite; E-extraction; Eop-extraction with oxygen and peroxide; C-chlorine; W-washing;
b ASB - Aerated Stabilization Basin; AS - Activated Sludge
c ADTM - Air Dried Metric Tonne
d Response - Where 7 d, 28 d, 56 d and complete life cycle tests were run, the range of concentration causing changes in mortality and growth are given, not the values for each bioassay.
e IC25 - Effluent concentration causing a 25% change in response; where the IC25 was not available, the lowest concentration causing effects was listed.
The toxicity of secondary-treated effluents from two bleached kraft mills was tested after introduction of oxygen delignification before bleaching by 60-70% ClO2 (Hall et al., 1996). Toxicity to various marine and freshwater invertebrates and fish was low, both before and after the process changes, so there appeared to be little change in toxicity, despite large changes in the discharge of chlorophenols and AOX.
The toxic effects found in the "before" and "after" studies by PAPRICAN and NCASI are similar to those of two other studies, one a mill before conversion to ECF, and one after conversion. Delayed sexual maturity and reduced egg production were found in fathead minnows exposed to secondary-treated effluent from a kraft mill using 50% substitution (Robinson, 1994; Table 4-2), with effluent toxicity similar to that of the pre-conversion study by NCASI (1985) and intermediate between the pre- and post-conversion studies by Kovacs et al., (1995a; 1996). While there seems to be general agreement among these studies of effects, toxicity, and the benefits of mill improvements, there is still variability among studies with respect to the concentrations of effluents causing specific effects. In fact, the concentrations cawing reproductive impacts on fathead minnows in life-cycle tests at a mill before chlorine dioxide substitution (NCASI 1985; Table 4-2) appear little different than concentrations of effluent from another mill after chlorine dioxide substitution (NCASI 1997, Table 4-2). However, as explained in NCASI (1997), much of this difference can be accounted for by water use. As indicated in Table 4-2, water use per ADMT of pulp varied by 2.5 fold among mills studied, and when taken into account, the concentrations of effluent after conversion appear less toxic than before conversion, even when comparing among mills. The improvements in comparability among studies after applying a simple normalizing parameter such as water use indicates that other normalizing techniques should be attempted. After normalizing to water use, there is also quite good agreement for results of studies of three ECF mills (Kovacs et al 1996; NCASI 1996; NCASI 1997; Table 4-2).
Compared to ECF effluents, less toxicity was observed with a secondary-treated TMP effluent (Kovacs et al., 1995a; Table 2). The mill produced about 600 T/d of hydrosulfite-bleached pulp from mixed western softwoods (predominantly white and black spruce) and about 8% recycled fibre. Effluent was clarified, passed through an equalization basin (9-12 h retention) and a secondary activated-sludge plant with 2-2.5 d retention. The effluent was not acutely toxic; i.e., LC50s for fathead minnow and Ceriodaphnia exceeded 100% effluent. Similarly there were no effects on 7-d growth and survival of fathead minnow larvae and only slight effects on 7 d Ceriodaphnia reproduction (1 sample in 4 caused effects at 65% (v/v)). During 200 d fathead minnow life cycle tests, there were no effects on mortality and growth (eggs through adults) at the maximum concentration tested (20% v/v), nor on sex ratios, time to maturity, reproductive activity or production of eggs. Samples taken for histopathology showed no unusual lesions. It was obvious from these tests that the "quality" of TMP effluent was greater than that of the kraft effluents. However, the reasons for higher quality are difficult to discern, because there were so many differences among the mills in technology and operation.
A battery of toxicity tests were applied by Ahtiainen et al. (1996) to whole effluents from mills using TCF, ECF and conventional chlorine bleaching; effluent was tested before and after secondary treatment. The effluents were derived from two BKMs bleaching hardwood and softwood pulps. Secondary treated effluents were non-toxic to bacteria (2 species), algae, Daphnia magna, and zebra danio embryos and larvae. Untreated effluents were sometimes toxic, with no differences in toxicity to Daphnia magna among the effluent types. Similar results were observed with the embryo-larval assay, although TCF and conventional untreated effluents from pine pulping were most toxic. Generally speaking, variability was so great that no clear conclusions emerged about the relative toxicity of the three effluents. This was true whether toxicity was expressed as percent effluent, toxic units or toxicity emission factors, although one untreated ECF effluent from hardwood pulping was exceptionally toxic. BOD, COD and TOC were all strongly correlated to toxicity (Versa et al., 1996), and correlation's of toxicity to concentrations of total phenols, fatty acids and resin acids could explain 79% of the variation in toxicity; AOX was less well correlated to toxicity. The conclusion of this study was that natural wood constituents accounted for most of the toxicity and that their concentrations were unrelated to ECF or TCF bleaching.
Ten of 15 bleached kraft mills studied by Priha (1996) used ECF bleaching and all employed some type of treatment. The exceptions to the ECF bleaching were TMP (N=1), conventional bleaching (N=1), combined ECF and TCF (N=2), and combined ECF and conventional bleaching (N=1). None of the mill effluents were toxic to Daphnia, and all secondary-treated effluents were non-lethal to zebrafish embryos, in contrast to effluents receiving only primary treatment. Only two of the effluents caused EROD induction in isolated rainbow trout cells, but seven were inhibitory. Correlations of toxicity to bleaching technology were not possible because none of the effluents was particularly toxic.
Recent data by Lovblad et al., (1994) on the toxicity of ECF and TCF effluents to MicrotoxTM, Selenastrum and Zebrafish (Danio rerio) suggested that toxicity emission factors were significantly reduced by TCF relative to ECF bleaching. However these conclusions were unjustified because all effluents had very low toxicity. Because toxicity was close to or greater than 100% effluent, relative toxicity, based on toxicity emission factors, was a measure of effluent flow rather than toxicity.
In summary, the primary ways to reduce effluent effects appear to be the prevention of black liquor losses and the provision of adequate secondary treatment. Despite the numerous studies cited above, there is little evidence that ECF bleaching is responsible for effects on fish. Some of the variation in toxicity among mills, or after mill improvements, might be more understandable if water use (ie., dilution of chemical constituents) was normalized by reference to factors such as COD, BOD, TSS and TOC (e.g., Willams et al., 1996). Expressing induction potency or toxicity in mass loading terms (e.g., Schnell et al., 1993; Lovblad et al., 1994) might also be a better indicator of which processes release inducers or toxicants, how much is released, and whether they are recovered during waste recycling and effluent treatment (Hodson et al., 1996b). The likely transfer of inducers to particulates and sludges during treatment means that a mass-balance based on both liquids and solids must be calculated for mill-wide or technology-specific comparisons, and this principle would also apply to substances causing reproductive effects.
This review has focused on the relationship between pulp mill technologies and effects of pulp mill effluents, with little regard for the chemical composition of those effluents. Previous reviews (Solomon et al., 1993; Hodson et al., 1996a) have demonstrated that effects cannot be related to AOX, and Williams et al. (1996) have shown the benefit of normalizing effects to DOC concentrations to account for dilution. Beyond these relationships, there are no measurements of effluent chemistry reported in studies of effects that provide a perspective on the role of technology. Concentrations of BOD, COD, TSS, phenols and resin acids are unrelated to the observed effects. With the developing understanding of the mechanisms of effects, and of the compounds associated with effects (e.g., retene, sitosterol, juvabione), future studies can employ more meaningful measurements of effluent chemistry. In fact, the study by Martel et al. (1996b) demonstrated clearly the close association between MFO induction and juvabione concentrations in different effluents of one pulp mill.
Studies of effluent toxicity to fish and invertebrates, and of MFO induction in fish, suggest that pulping is the most important source of substances causing toxicity. These would likely be wood extractives or breakdown products of lignin. While there are clear benefits of black liquor spill control and of secondary treatment, no studies demonstrated environmental advantages of TCF bleaching compared to ECF bleaching There were also no studies that established a clear link between toxicity reduction and the installation of any other specific technologies. Improvements in mill effluents were associated with multiple changes to many aspects of mill operations, and could not be ascribed to any one technology. Future advances in this area rely on a mass balance approach that accounts for both liquid and solid wastes and on the measurement of the compounds associated with effects.
Cates et al. (1995) used lab-scale bleaching of hardwood and softwood pulp (TCF sequence, pine: xylanase (X), DTPA (Q), Ep, Ozone (Z) and peroxide (P); eucalyptus: XZQP) (ECF, hardwood and softwood: DEDED). Bioassays of 15-minute Microtox at pH 6.3 gave EC50s ranging from 0.24% effluent (ECF) to 0.56% (TCF). Treatment of effluent with white rot fungus reduced toxicity to >1.0% effluent, the highest concentration tested. Consequently, the effect of bleaching on the toxicity of final treated effluent was unknown.
O'Connor and Nelson (1996) tested the toxicity of effluents from laboratory bleaching of pulps using ECF (OD1OOEop) and TCF (OQP) sequences. Bioassays included acute lethality to Microtox, Ceriodaphnia, and Daphnia and reproductive effects on Ceriodaphnia. The acute lethality of the ECF lab bleach effluent was similar to that of a real ECF mill effluent (100 and 87% v/v respectively), as were reproductive toxicity thresholds (53 and 56% v/v respectively). The acute lethality of TCF bleaching effluents from four wood species varied considerably. Although non-toxic to microtox bacteria, the LC50s for Daphnia and Ceriodaphnia varied from about 50% to >100% (v/v), depending on whether cedar (most toxic), hemlock, douglas fir, or spruce/pine/fir (least toxic) were being tested. In all cases, effluents were quenched with metabisulfite to avoid bias due to residual peroxide. These effluents were also toxic to Ceriodaphnia reproduction, with threshold concentrations ranging from 4.7 to 13% v/v; the differences among wood species were less pronounced. By comparison, the spruce/pine/fir effluent from ECF bleaching was about 10-fold less toxic (53% v/v) than effluent from TCF bleaching (5.1% v/v). These tests were of bleaching effluent only and did not determine whether treatability to remove toxicity in final effluent would vary with the type of bleaching used.
Effluents generated in the laboratory from ECF bleaching sequences were either equally toxic or less toxic than effluents from TCF bleaching sequences. However, because experimental designs were incomplete, the effect of bleaching on treatability, and hence on the toxicity of final effluents, was not evaluated. Studies since 1993 have identified natural compounds causing MFO infection and impaired regulation of steroid hormones in fish. However, few studies directly addressed the question of the role of bleaching technologies. There was no compelling evidence that effluents from pulp mills using ECF bleaching are any more or less toxic, or cause any more or less MFO induction than those from mills using TCF bleaching. Progress in this area requires a mass balance approach and chemical analysis of effluents to understand relative toxicity's of effluents from different mills and technological processes. The importance of sludges, particulates, and dissolved organic carbon as a source of compounds affecting fish has not yet been properly evaluated.