Sampling Success
A facility learns that the method used to determine the pH of an anaerobic digester can significantly affect the value obtained
In 1986, the King County Department of Natural Resources and Parks (Seattle) expanded the capacity of its South Treatment Plant (formerly known as the Renton Treatment Plant) to an average wet weather design capacity of 272,520 m^sup 3^/d (72 mgd). As part of that effort, facilities for processing solids were constructed, including dissolved-air flotation (DAF) systems to thicken primary and secondary solids, anaerobic mesophilic digesters, and belt- filter presses for dewatering biosolids. Before this expansion, the facility transferred its raw solids to the county’s other treatment plant via a 4.8-km (3-mi) force main for further treatment.
Startup of the solids-processing systems began in November 1987, and the anaerobic digesters came on-line with minimal operational problems. Digester sludge was sampled and monitored daily for a number of parameters, including total solids, volatile solids, alkalinity, volatile acids, ammonia, and pH. Temperature and gas production were monitored using on-line instrumentation, and gas samples were analyzed daily for carbon dioxide. Once startup was completed, the digesters settled into a stable operating mode with steady gas production and composition. Volatile acids consistently were under 100 mg/L, and alkalinity consistently remained above 2000 mg/L.
Although digester startup and operation proceeded smoothly, the belt-filter presses produced dewatered biosolids at a much lower solids concentration than expected. Troubleshooting efforts initially focused on operational factors associated with belt- filter press performance. A number of improvements relating to polymer mixing and solids loading rates were identified, helping to improve the dewatering process.
Additional efforts to improve dewatering focused on the high proportion of secondary solids in the digesters. Because secondary solids are harder to digest and dewater than primary sludge, staff proposed that a longer digester solids retention time (SRT) might improve dewatering performance. A full-scale study on the effects of digestion time on dewatering performance confirmed this theory. Subsequently, an additional digester was placed into service, increasing the digester SRT from 21 to 28 days.
In the meantime, startup and operation of the DAF thickeners proved troublesome. Problems controlling thickened solids concentrations jeopardized the ability to maintain the longer SRT needed for adequate digestion of secondary solids. Operations, process control, and maintenance staff refocused their efforts on optimizing the operation of the DAF thickeners. These efforts proved successful, resulting in steady improvement in the concentration of the thickened sludge (see Figure 1, p. 22) and, thus, in the digester feed concentration. The thicker sludge yielded a variety of benefits, including increased digester SRT, an increase in the concentration of digester sludge, reduced digester heating requirements, and an increase in digestion capacity. Together with the other process optimization work, these changes improved the dewatering performance of the belt-filter press operation.
As successful as the efforts were at improving thickening, the resulting high digester pH and ammonia concentrations prompted concern for ammonia toxicity. Figure 2 (see p. 22) shows how digester pH increased over time. Ammonia-nitrogen concentrations displayed a similar trend, reaching levels in the range of 1600 to 1800 mg/L.
Concern regarding potential ammonia toxicity intensified following a literature review. Several authors reported the development of toxic conditions in digesters with ammonia concentrations of 1600 to 1900 mg/L and pH levels between 7.7 and 7.9, conditions present in the county’s digesters. Although these “toxic” conditions were present in the facility’s digesters, the continued stable operation of the digesters tended to allay immediate concerns. Therefore, it was decided to continue to monitor digester performance and employ options such as adding acid to reduce the digester pH in the event that digester performance began to deteriorate. Fortunately, the digesters operated in a stable manner for many years.
Figure 1. Thickened Sludge Concentration Versus Time
In late 1993, work began on the predesign phase of an effort to increase the South Treatment Plant’s solids treatment capacity by 25%. In general, the resulting evaluation increased design loadings of certain unit processes and concurrently raised the design value for the thickened sludge concentration from 4.5% to 5.5%. Although process control staff determined that even higher concentrations could be achieved, several constraints, including issues related to mixing, pumping, and potential ammonia toxicity, prevented this change from being implemented. Concerns about possible adverse effects on digester mixing were considered, and it was decided that feed concentrations in excess of 7% would have to be reached before such problems became a real issue. Of greater concern was the potential to develop ammonia toxicity associated with the higher feed concentration and the concomitant increase in ammonia concentration at the measured pH. This concern prompted process control staff to evaluate the system closely.
Figure 2. Digester pH versus SRT
Table 1. Data From In Situ pH Measurement Versus Grab pH Measurement
Because the accuracy of the ammonia analysis was not in question, the logical variable upon which to focus was pH. After reviewing the literature and sampling procedures, the project team concluded that the method used to sample and analyze for pH might account for the discrepancy between measured pH and ammonia levels and actual digester performance. It was hypothesized that the release of carbon dioxide from samples during the period between collection and analysis might account for a rise in pH. Such a result would indicate the potential for ammonia toxicity, even though the in situ pH was low enough to ensure that the unionized ammonia concentration was safely below the reported toxic level.
Using Direct and Indirect Sampling Methods
Several approaches were used to evaluate the effect of carbon dioxide evolution on sample pH. The first approach was to insert a pH probe directly into a line containing digested circulating sludge to avoid potential off-gassing of carbon dioxide. A second approach involved measuring the pH of the digested sludge with a field pH meter while the sample was withdrawn. Specifically, a sample of circulating sludge was drawn directly from the circulating sludge line and allowed to overflow the sampling container into which the pH probe was inserted. Once a stable pH reading was obtained, the process of withdrawing the sample was stopped and any change in pH was observed over time as the sample was mixed.
In addition to these two direct analytical approaches, two indirect methods were evaluated for use in determining digester pH. The first indirect method calculated pH by measuring the concentrations of hydrogen sulfide in the digester gas and total soluble sulfide in the circulating digester sludge. The second indirect method attempted to estimate pH using a similar process based on gas-phase ammonia and solution-phase ammonia. These methods then calculated digester pH using the equilibrium constant for gas- liquid partitioning – in other words, Henry’s law – and the acid dissociation constant (K^sub a^).
Chemical analyses generally were performed according to procedures contained in Standard Methods for the Examination of Water and Wastewater. The pH was measured using either a laboratory or a portable pH meter depending on the specific application. Gas- phase concentrations for hydrogen sulfide and ammonia were obtained using adsorption tubes manufactured by Sensidyne Inc. (Clearwater, Fla.). Liquid-phase ammonia was measured using a specific ion electrode by the method of standard additions. Liquid-phase total dissolved hydrogen sulfide was measured using either a specific ion electrode by a method published by the Hach Co. (Loveland, Colo.) or the methylene blue method included in Standard Methods. Values for Henry’s law constant and the acid dissociation constants (K^sub a^) for hydrogen sulfide were obtained from Sulfide in Wastewater Collection and Treatment Systems, a manual published by the American Society of Civil Engineers (Reston, Va.).
Eliminating incorrect digester pH readings helped Seattle’s South Treatment Plant get the most from its belt-filter presses, increasing cake solids by nearly 5%.
Assessing the Results
The initial work was directed toward obtaining an in situ pH measurement by means of an insertion probe. Process staff and plant maintenance staff modified a portion of the circulating sludge line to accept an insertion probe. Concurrently, a grab sample was collected and transported to the plant’s process control laboratory where pH was measured (see Table 1, p. 23). Although these data confirmed the theory that actual pH values in the digesters were substantially lower than was indicated by the grab samples, they were difficult to obtain. Specifically, staff ha\d a hard time inserting and removing the pH probe without potentially subjecting themselves to a “sludge shower.” Therefore, this approach was discontinued once data were obtained confirming the hypothesis.
Table 2. Total Dissolved Sulfide Concentration Versus Gas Phase Hydrogen Sulfide Concentrations at Various pH Values
The second approach involved a procedure to measure pH immediately upon withdrawing a sample. This required the use of a portable pH meter and an overflow sample container into which the probe was inserted and the sample was discharged. A method similar to that used for collecting dissolved oxygen samples was used, meaning that a sample was discharged via a tube at the bottom of the container and allowed to overflow. The pH reading was taken while the discharge was occurring to minimize the potential for pH change from loss of carbon dioxide.
Despite every effort to minimize the loss of carbon dioxide, pH values obtained with this procedure were close to those obtained with the grab-sample method. For example, pH values ranged from 7.5 to 7.7 with this method, while the values obtained with the insertion probe ranged from 7.1 to 7.2. Therefore, it appears that the major loss of carbon dioxide occurs because of the pressure drop and turbulence across the sample valve. Although this approach was not pursued further, it may be possible to collect a sample in an enclosed container that can be transported to a laboratory for analysis without enabling off-gassing to occur.
Because ammonia was the constituent of primary concern, it was thought that a gas-liquid equilibrium approach similar to that proposed for hydrogen sulfide would be worth investigating. The digested sludge would be analyzed for ammonia, and, therefore, it would only be necessary to add the analysis of digester gas ammonia to complete the procedure. Although the odor of ammonia was readily evident in the dewatering area, attempts to measure ammonia concentrations in digester gas proved unsuccessful. The concentration was significantly below the method detection limit of 5 ppm by volume. Based on equilibrium calculations, the gas-phase ammonia concentration should be at or below the analytical detection limit at a digester pH of 7.2. Although options to concentrate ammonia were considered, they never were implemented.
The final approach involved measuring concentrations of gas- phase and liquid-phase hydrogen sulfide. Concentrations of hydrogen sulfide in digester gas were measured routinely and consistently found to be in the range of 50 to 200 ppm. Samples of circulating sludge were analyzed for total soluble sulfide and found to range from 0.1 to 0.3 mg/L. This procedure’s accuracy depends on the potential for loss of hydrogen sulfide during sample preparation, and, thus, actual concentrations could have been somewhat higher than measured.
Discerning the Actual pH
The results obtained using the insertion electrode substantiated the assumption that digester pH was significantly lower than the grab-sample analysis indicated. Insertion values were in the range of 7.0 to 7.2, levels normally expected in well-operated digesters. While this procedure accurately indicates digester pH, the value of on-line measurement probably would not be justified given the potential fouling of any device inserted in a circulating sludge line. Although the county’s current procedure is to continue collecting grab samples, employees are investigating procedures that would enable in situ data collection.
Figure 3. Gas Phase H^sub 2^S vs Dissolved Sulfide
Measuring pH while withdrawing samples directly from the circulating sludge line resulted in values comparable to those obtained using the grab-sample procedure. Unless some method can be developed to prevent the immediate release of carbon dioxide, this method does not offer any advantages over the current procedure. Although the value obtained using the continuous overflow method was slightly lower than that from the grab sample, it rapidly approached that value over time.
Extracted from Sulfide in Wastewater Collection and Treatment Systems, these equations, when combined, generated the concentrations of total dissolved sulfide versus gas-phase hydrogen sulfide concentrations at various pH values (see Table 2, p. 24).
As noted earlier, soluble sulfide concentrations are generally in the range of 0.1 to 0.3 mg/L in the South Treatment Plant’s digesters, while gas-phase sulfide concentrations are usually between 50 and 200 ppm. As a result, the data in Table 1 were converted to a nomograph (see Figure 3, p. 25) that covers a low range of soluble sulfide concentrations. In areas in which higher sulfide levels are encountered, the nomograph, the table, or both can be expanded as needed. As can be seen from Table 1 and Figure 3, the total dissolved sulfide value and the atmospheric concentration of hydrogen sulfide also support the fact that the actual pH within the digester is in the range of 7.2 rather than the values of 7.7 to 7.9 reported from the grab-sample analysis.
Benefiting From Better Data
Although the variable accuracy of the method used to measure digester pH may matter little to facilities that are not loading their digesters to the level occurring at the South Treatment Plant, the increased demand by the public to maximize the use of existing facilities ultimately will result in more facilities experiencing these same conditions. As was the case at the South Treatment Plant, incorrect pH readings can lead to a decision not to push the thickening process, and thus the digesters, to the extent to which they are capable. This information has enabled staff at the facility to further improve the operation of the DAF thickeners to the point where digester feed solids are approaching 6.5% to 7.0% without causing concern about potential ammonia toxicity.
In addition, the South Treatment Plant demonstrated an ability to increase SRT and improve biosolids dewaterability. The steady increase in SRT has been accompanied by a stable increase in cake solids produced by the belt-filter presses. The combination of efforts such as the ongoing increase in digester feed concentration and SRT led in part to an increase in cake solids, from 17% to almost 22%. The result was a significant savings in the cost to haul and apply the biosolids and, thus, a significant savings to the agency. The facility’s belt-filter presses were replaced with high- solids centrifuges in 2005.
In the past, digester pH primarily has been used to measure digester stability, and a drop in pH was cause for major concern. Under that scenario, the absolute value of the pH was of less concern than the change in pH. With the increasing emphasis on maximizing the use of process units, factors such as potential ammonia toxicity in the digesters will become a similar concern. Even if steps are not taken to address the issue of elevated pH readings caused by off-gassing, recognizing the phenomenon and verifying periodically the in situ pH using any of the methods discussed above will indicate with assurance whether ammonia toxicity is of concern and if further improvements in the thickening process or other steps to maximize digester operation can be continued or should be avoided.
For More Information
American Society of Civil Engineers (1989). Sulfide in Wastewater Collection and Treatment Systems. Reston, Va.: ASCE Manuals and Reports Engineering Practice, No 69.
Richard E. Finger is west section manager with the King County Water Pollution Control Division’s West Point Treatment Plant (Seattle). Richard C. Butler is process control supervisor with the King County Water Pollution Control Division’s South Treatment Plant (Renton, Wash.).
Copyright Water Environment Federation Aug 2005