Everett benefits from teamwork

An effort worthy of Sherlock Holmes solved the mystery of how this Kimberly-Clark mill was producing elevated levels of furans despite ECF bleaching

December 2007
By Christine Kurtz, Kenneth S. Weiner and Doug Asbe

When Kimberly-Clark (K-C) converted its Everett, WA, mill to elemental chlorine-free (ECF) bleaching technology, in part to reduce chlorinated organics as required by federal law, the last thing it expected to find was elevated levels of chlorinated furans (furan). After all, controlling furans is one of the drivers behind the global trend away from chlorine bleaching.

As a result, the Everett mill embarked on a six-year sleuthing effort from 2001 to 2007 worthy of Sherlock Holmes. In the end, the mystery was solved through cooperation between K-C, the City of Everett, the Washington State Department of Ecology, and the community. The mill, the city, public and worker risk – and Puget Sound, one of America's prized estuaries – all benefited.

Sunrise over Kimberly-Clark's Everett mill where furans were posing a problem.

ECF Surprise

K-C's Everett mill includes a pulp mill that was built in 1931 and produces 150,000 tons/yr of ammonia-based acid sulfite pulp. The pulp mill has been upgraded several times, most recently in 2000 with the elimination of elemental chlorine and sodium hypochlorite for pulp bleaching and the addition of a Solvay chlorine dioxide generator and two chlorine dioxide bleach stages. More than half of the pulp produced is used on-site to make various tissue products. The remaining pulp is mostly sent to other K-C tissue facilities.

The Environmental Protection Agency (EPA) Cluster Rule (40 CFR 430.54) mandated that ammonia-based sulfite pulp mills install ECF bleaching technology by December 31, 2000. The Cluster Rule included a proposed 10 parts per quadrillion (ppq) concentration limit on 2,3,7,8-tetrachlorodibenzofuran (furan) in bleach effluent. Whether or not this limit is appropriate for ammonia-based sulfite mills is arguable (kraft pulp mills have a limit of 31.9 ppq), but mill personnel initially did not foresee any issues with meeting the proposed limit.

Previous studies indicated wood-based furan precursors were destroyed by the sulfite cooking process. Other sources of furan precursors, such as defoamers added to the process, were reformulated to reduce the precursors. With precursors greatly reduced and the ECF conversion, there should have been no measurable furans.

Initial testing of bleach plant effluent in 2001, however, showed the mill might not be able to meet the proposed limit, even with the installation of the required ECF technology. This result was unexpected.

Overview

In September 2002, the mill began a comprehensive study to determine all sources of furan in the bleach effluent and to develop a viable control strategy. The initial data from 2001 and 2002 indicated higher furan concentrations occurred during the summer months. Subsequent testing through 2007 shows the same seasonal variation.

Two significant sources of furan were discovered. Both are the result of natural, non-chlorinated furan precursors contained in the incoming water supplied to the mill. The first significant furan source was the water itself. The City of Everett provides the mill with water from a high quality mountain reservoir in a dedicated line to the mill ("incoming water"); separate from the city's potable drinking water. Furans were found in the water before it entered the mill. Chlorine, added by the City of Everett to the water as a disinfectant to control biological growth in the transport pipelines to the mill, reacted with the furan precursors in the water to form furan.

The second significant furan source was found to be the chlorine dioxide solution. The Solvay generator produced low levels of chlorine as a byproduct. This chlorine reacted with the precursors in the water to form furan in the chlorine dioxide storage tanks.

Through the process described here, K-C determined the source of the furan and developed specific process changes to reduce furan concentration. To address furan in the chlorine dioxide solution, K-C converted its chlorine dioxide generator to use hydrogen peroxide instead of methanol as a reducing agent, thus reducing the amount of byproduct chlorine available to react with the raw water precursors. To address furan in the incoming water supplied to the mill, the City of Everett replaced chlorine with sodium hypochlorite for water disinfection, again reducing the amount of chlorine available to react with the precursors. State regulators supported the mill's effort to track down and solve the problem with good science.

These combined efforts resulted in controlling the unexpected generation of chlorinated furan in both the City's water supply to the mill and in the mill's bleach plant effluent, which discharges to Puget Sound.

Fig. 1 - Dibenzofuran

What is Furan?

Dibenzofuran is a heterocyclic organic compound with the structure shown in Fig. 1.

It is naturally occurring and can be found in some wood species. Under certain conditions, dibenzofuran will react with chlorine and chlorine compounds. The degree of substitution depends on the method of chlorination. Direct substitution can occur in the presence of elemental chlorine. Chloride salts will combine with dibenzofuran when heat is applied. This reaction can happen naturally, as in a forest fire, or artificially, as in a furnace.

Chlorinated dibenzofuran compounds are considered toxic and are chemically stable. They do not break down easily and accumulate in fatty tissues. The EPA considers furan a persistent, bioaccumulative and toxic (PBT) chemical. One of the concerns with PBTs in the marine environment is their tendency to concentrate in the tissue of long-lived predators at the top of the food web, such as salmon, seals and whales.

Control of PBTs has taken on added urgency in many places as a threat to the health of humans and other species. For example, Puget Sound salmon and orcas (killer whales) are both currently listed as endangered species. There are many causes, and source control of toxicants is part of the recovery strategy.

The EPA chose to regulate levels of 2,3,7,8 tetrachlorodibenzofuran (furan) with the structure shown in fig. 2, as part of the Cluster Rule. This isomer is considered the most toxic of the furans and is primarily formed by direct chlorine substitution. Chlorine dioxide does not produce chlorinated furan compounds. Research at the time the Cluster Rule regulations were written indicated ECF bleached ammonia-based sulfite mills would not produce furans. The regulated limit was thus proposed to be 10 ppq (the detection limit at that time).

Today, analytical testing for furan is remarkably accurate from some labs at two parts per quadrillion (ppq).

Fig. 2 - 2,3,7,8 tetrachlorodibenzofuran

Comprehensive Study

Initial reviews of the data showed seasonal variations in furan, but were inconclusive to the source. The mill started a comprehensive study in September 2002. Samples of each bleach stage effluent (as well as several raw materials) were taken and tested. Table 1 shows the results from the initial round of testing.

The highest levels of furan were detected in the EoP effluent, which was expected. Any furans created in the first chlorine dioxide stage would become soluble in the caustic conditions of the EoP tower. The furan concentration in the incoming water was not expected.

Seasonal Variations

Routine monthly sampling of the incoming water began in October 2002. By July 2003 a pattern seemed to emerge. High concentrations of furan in the incoming water occurred in the summer months - matching the pattern observed in the bleach plant effluent (see Fig. 3). This pattern was confirmed as the study continued.

Figure 3 shows the monthly average furan concentration for the bleach plant effluent and the incoming water. Summer months have higher furan concentrations. The bleach plant effluent data are separated by the period before the chlorine dioxide generator conversion (2001-2003) and after the conversion (2004-2007). Note that since the conversion, there have been no detectable furans in the months of November or December.

The mill's National Permit Discharge Elimination System (NPDES) wastewater permit was due for renewal at the time the furans were detected. Mill engineers began voluntary updates on the furan study with the Washington State Department of Ecology (Ecology) in December 2002. The mill planned to expand the study the following summer to run trials on technologies to control the generation of furans at their source, both in the water supply and in the mill.

In 2004, Ecology was tasked with renewing the mill's new NPDES permit including a limit on furans while also allowing time to complete the planned trial work. There was an administrative appeal of Ecology's renewal of the permit since the furan issue was not yet solved. The dilemma was that the trials needed to occur over several years during the warm season – when the furans were generated – to allow variables to be isolated and tested scientifically. The time was constricted further by the four-week lag time between sampling and analyzing results.

K-C and the interested parties reviewed the study plan together. They agreed on a timely, scientific approach under a three-year compliance schedule (2004-2007) in the renewed permit. As the three-year study progressed, mill personnel periodically met with interested parties and kept them informed of progress with regular reports.

Fig. 3 - Monthly average furan concentration

Furan in Incoming Water

Since the mill operations are relatively constant year-round, the seasonal variations suggested that the source of furan generation might be outside of the mill. Incoming water turbidity and temperature were known to vary seasonally and were included in the initial furan tests. Once furan was detected, engineers from K-C worked with the staff at the City of Everett water treatment plant to identify the source of furan.

Incoming water to the mill is supplied by the City of Everett through a dedicated water line. The process water is drawn directly from a city-owned source and is supplied to the mill following chlorine disinfection. The chlorine disinfection is accomplished using a small stream of concentrated chlorine and raw process water. City staff adjust the chlorination solution feed rate to maintain a 1 ppm free chlorine residual in all process water supplied to the mill.

This concentrated feed line was tested for furan on several occasions, as shown in Table 2. Furan concentrations were more than 1,000 ppq. These results encouraged the City of Everett and K-C to look for alternatives to chlorine disinfection, both for the municipal and the mill water supplies.

Furan in Mill Process

Incoming water furans alone did not account for all of the furan in the bleach plant effluent. A series of trials were performed with several bleaching and cooking conditions aimed at identifying where furans were formed in the pulp mill.

Ultimately, bleaching and pulping conditions were shown to have little correlation with the measured furan concentrations in the bleach plant effluent. Figures 4 through 9 show furan concentration as a function of:

• Pulp kappa from the digesters, a measurement of lignin content, Fig. 4;
• Chlorine dioxide dosage to the D1 stage, Fig. 5;
• Wood species: hardwood or softwood, Fig. 6;
• Salt (NaCl) content in sodium chlorate, Fig. 7;
• Chlorine dioxide dosage to the D2 stage, Fig 8;
• Incoming water temperature, Fig. 9.

Fig. 4 - Furan concentration as a function of pulp kappa

Pulp kappa, D1 tower dosage and wood species show little if any correlation with furan concentration. Contrary to expectations, high ClO₂ addition rates in the second D stage and high salt in the sodium chlorate result in lower concentrations of furans.

Only warmer incoming water in the summer results in more furans. As water temperature increases in the summer, Fig. 9 shows higher furans with warmer summer water conditions. This supports the seasonal variations observed in Fig. 3 and Table 2.

Fig. 5 - Furan concentration as a function of ClO2 dosage to D1

Fig. 6 - Furan concentration as a function of wood species

Fig. 7 - Furan concentration as a function of salt content in chlorate

Chlorine Dioxide Reactors

The mill's chlorine dioxide storage tanks provided the conditions necessary to generate furan. Some styles of reactors such as the older Solvay chlorine dioxide generators produce elemental chlorine in a side reaction. Chlorine concentration in the chlorine dioxide solution at the Everett mill averaged 0.5 g/L. This was apparently high enough to chlorinate the furan precursors.

Chlorine dioxide solution was tested for furan on several occasions between November 2003 and August 2004. These results are shown in Table 3.

Concentrations up to 288 ppq were observed. These results accounted for half of the furans in the bleach plant effluent.

Solutions

With two significant sources of furan identified, K-C mill engineers and City of Everett staff scientists, with the help of scientists at the Pulp and Paper Institute of Canada (Paprican), identified technologies that might reduce furan levels. These included actions by K-C and the City of Everett. The candidate technologies and trials included the following:

pH: In the spring of 2004 the pH of the chlorine dioxide solution was raised to 7 with sodium hydroxide. Direct chlorination of furan does not occur at neutral pH levels. This trial was not effective and caused corrosion and operational problems.

No chlorination: In August 2003 the City of Everett stopped chlorinating the mill water line for one month to measure the effect of eliminating precursor chlorination. This action resulted in lower furan levels in the water and in the bleach effluent.

Hydrogen peroxide: In June 2004 the chlorine dioxide generator was converted to use hydrogen peroxide instead of methanol as a reducing agent. This technology was effective and reduced furan levels by half. Residual chlorine concentrations after conversion averaged 0.2 g/L.

Activated charcoal: Between May 2004 and November 2005, activated charcoal was used to filter out the furan precursors in the water used for chlorine dioxide solution manufacture. The filters did not prove to be effective.

Sodium hypochlorite: During the summer of 2004 the City of Everett replaced chlorine with sodium hypochlorite for water disinfection during two trial periods. These trial periods were effective at reducing furan in the water and the bleach effluent. The trial was repeated in July and August of 2007 and again was effective at reducing furan.

Fig. 8 - Furan concentration as a function of ClO2 dosage to D2

Fig. 9 - Furan concentration as a function of incoming water temperature.

Results

Because furan generation was occurring mostly under warmer conditions – which are seasonal in the temperate Pacific Northwest – it took three years to run the trials, and two succeeding summers to confirm the control technologies were effective.

The chlorine dioxide generator conversion successfully reduced the furan generated in the mill's bleach plant effluent. In 23 of the 30 months between June 2004 when the conversion was completed and November of 2006, the bleach effluent furan concentration was actually less than or equal to the incoming water.

In an effort to eliminate the safety issues associated with chlorine gas, the City of Everett began construction of new disinfection facilities that use sodium hypochlorite. During the summer of 2007, pending completion of the new facilities, the City implemented a temporary sodium hypochlorite disinfection system for the process water supplied to the mill. The new permanent hypochlorite facilities will be completed by the spring of 2008.

The combination of the two process changes has resulted in the first "non detect" furan concentrations in the bleach effluent for the warmer months of July and August since sampling began in 2000.

Conclusions

The existence of furan precursors in incoming water, along with the effects of heat and chlorine, are conditions that could occur at facilities other than ammonia-based sulfite pulp mills. Furan generation was unexpected in light of the mill's conversion to ECF bleaching and its ammonia-based sulfite pulping process. Finding the furan sources and the technological solutions to remove them required a phased scientific study. This study involved K-C, the City of Everett, State of Washington regulators and citizen groups. The resulting control measures are relatively basic. Since implementation, furan concentrations in the mill's bleach plant effluent have consistently been lower than the10 ppq permit limit.

The combined efforts of K-C and the City of Everett produced benefits beyond improving the mill's wastewater discharge. K-C was able to eliminate the use of methanol, a highly flammable liquid.

The City of Everett's actions to eliminate gaseous chlorine, an inhalation hazard, and an EPA Environmental Risk Management compound, allowed the mill to realize the side benefit of reduced furan formation.

"The City has a long history of collaboration with the mill on technical issues," states Robert Waddle, the City's operations superintendent. "Working through the furan formation issue with K-C staff is just the latest example of our public-private partnering."

"It would have been easy for everyone to point the finger at someone else," says Chris Isenberg, Everett mill manager and pulp mill manager at the time. "Instead, everyone worked together to find the sources, which wasn't an easy task, and to control them.

"It's a good result for the mill, the environment and public health. I hope we can use this kind of deliberative, cooperative approach more often in tackling elusive problems," he adds.

Christine Kurtz is environmental and safety manager, K-C, Everett mill; Kenneth S. Weiner is partner, K&L Gates LLP, Seattle, WA; Doug Asba is senior environmental/process engineer, Bellingham, WA.

Pulp & Paper magazine is FREE to qualified subscribers. Click here to find out more.