membrane forensics and troubleshooting membrane ...

Presented at the 2006 AMTA Biennial Conference in Anaheim, CA on August 1, 2006

MEMBRANE FORENSICS AND TROUBLESHOOTING MEMBRANE FACILITIES Steven J. Duranceau, PhD, PE, Boyle Engineering Corporation, Orlando, FL 32801

ABSTRACT The purpose of this paper is to: Provide an overview of membrane processes and fouling, describe biofouling as a biofilm problem, and examine countermeasures; Discuss membrane forensic procedures and describe basic autopsy testing practices; Describe appropriate countermeasures and prevention options available to membrane plant designers and operators to combat and prevent fouling. OVERVIEW OF MEMBRANE PROCESSES Understanding membrane application requires understanding of the characteristics of drinking water membrane processes. Reverse osmosis (RO), nanofiltration (NF), electro-dialysis reversal (EDR), ultrafiltration (UF) and microfiltration (MF) are the membrane processes, which have application to drinking water. Combinations of membrane processes with other processes have become known as integrated membrane systems. Although a conventional NF process consists of a pretreatment and post treatment process before and after the NF, which could be described as integrated, this is described as conventional. The coupling of a MF and a NF or coagulation, sedimentation and filtration with a NF are accepted examples of an integrated membrane system (IMS). The basic characteristics of these processes are shown in Table 1. Although many factors affect the solute separation by these process, a general understanding of application can be achieved by associating minimum size of solute rejection with membrane process and regulated contaminate. Table 1. Characteristics of Membrane Processes Process EDR RO NF UF MF Mechanism: Pathogens: Organics:

Mechanism C S, D S, D S S

Exclusion 0.0001 μ 0.0001 μ 0.001 μ 0.001 μ 0.01 μ

Regulated Solutes Pathogens

Organics

Inorganics

None C, B, V C, B, V C, B, V C, B

None DBPs, SOCs DBPs, SOCs None None

All All All None None

C=charge, S=sieving, D=diffusion C=cysts, B=bacteria, V=viruses DBPs=disinfection by-product precursors, SOCs=Synthetic Organic Compounds

MEMBRANE FOULING There are four primary mechanisms of membrane fouling: (1) Plugging and/or entrainment; (2) Scaling; (3) Organic adsorption; and, (4) Microbiological Scaling control is required for all RO/NF membrane systems in either surface or groundwaters and is achieved by acid and/or antiscalent addition. Plugging control is required for all RO/NF membrane systems in either surface or groundwaters and is achieved by feed water turbidities and silt density indexes (SDI’s) less than 0.2 and 2 respectively. Bio-fouling control is typically required for aerobic surface or groundwaters and is achieved by monochloramine (NH2Cl) or addition of other bactericidal agents. Organic fouling can occur in surface water systems with TOC > 3-6 mg/L and is typically reduced by coagulation, sedimentation and filtration or advanced pretreatment. Water quality will certainly be the determining factor on type of membrane processes utilized in the treatment process and design. Surface water supplies will in the future more readily employ integrated membrane processes, and can include both fresh, brackish and seawater sources. River water supplies will represent the most variable water quality, particularly in terms of particle loadings and turbidity. Consequently, control of membrane fouling for surface water systems can be significant. THE BIOFILM PROBLEM One of the more common problems currently impacting membrane process operations today includes inadequate microbial control, often referred to as biofouling. In order for bacteria to grow rapidly and extensively to cause serious problems in a membrane process requires that certain conditions be satisfied: (1) the presence of food; (2) a warm environment; (3) flowing water; and, (4) the absence of disinfectant. A single viable microorganism replicating every 20 minutes under ideal conditions can produce progeny in excess of 4.7 x 1021 within one day. Therein lies the nature of biofouling of membrane processes. The biofouling organisms of most concern to designers and operators of membrane facilities are those that upon adhering to a surface will excrete a protective mass of polysaccharides known as extra-cellular polysaccharide substances. The biofilm serves to entrap nutrients and can block feed channel passages resulting in differential pressures across multielement pressure vessels in excess of manufacturer recommended pressures. Increased differential pressures can permanently damage membranes and overall increase operating costs. Bottom line is that biofouling leads to considerable technical problems and economic loss.

MEMBRANE AUTOPSIES It is understood that the most important component of a membrane system, the membrane, contains a significant amount of information on an RO’s operating history once an RO plant is placed on-line. Aged membranes can be systematically evaluated to determine the type of foulants; these foulants and ancillary autopsy information could be used to provide information that impacts other operating systems and help resolve those problems, if any exist. Prior to dissection of the membranes, most autopsies begin with visual and wet testing examinations of the exterior of the membrane elements received. The autopsy typically will document if the membrane elements received exhibit a stagnant biological odor upon analysis. In addition, if the feed channel spacer material has been noted to extrude from the membrane concentrate end, and indication of extreme pressure drop and initial telescoping effects. After element characterization, inspection and foulant collection, a loss-on-ignition (LOI) test followed by targeted energy dispersive x-ray (T-EDXA) and a surface backscatter analysis using fourier transform infrared (FTIR) spectroscopy is conducted. Other common analysis that are useful in understanding membrane process problems include: microscopic analysis, scanning electron microscopy (SEM), biological testing and cell testing procedures. Wet testing refers to an evaluation of the membrane elements performance where the membrane is tested in a single element pressure vessel under laboratory conditions. Pressure drops, salt rejection (based on a feed solution of 1000 mg/L sodium chloride), temperature and flux rates are measured and normalized then compared to manufacturer warranty conditions. The membrane elements tested as received are usually evaluated in terms of a specific salt rejection (for example, 98.3%), a permeate flow (say for example 8226 gallons per day (GPD)) and a pressure drop (for example 15 psi). This compares to a manufacturer specification for this specific example of 7700 GPD at 98.6% nominal 97.0% minimum rejection (brackish water) and maximum pressure drop of 10 psi. Obviously the membrane element tested failed for pressure drop criteria. One typically will take the wet-tested membrane and perform destructive autopsies of the membrane surface. Dissection of the element evaluated often reveals a moderate coating of a tan colored gelatinous foulant that uniformly covers the membrane surface of all of the element’s membrane leaves when biofouling has occurred. The edges of the feed-end of the membrane may also contain a black foulant. Often the membrane element is evaluated microscopically using magnification ranges between 50X and 1000X. Microsopic data indicated that the source (feed) water is often highly variable with respect to microorganisms when biofouling occurs, to the extent that sometimes strains of non-flagellated algae can be detected in the lead element. Often if pretreatment failed, or contractor debris is present during startup, then membrane fouling may occur. Figure 1 displays an example of the results of LOI testing of a membrane element that had fouled. LOI is conducted by first taking foulant sample and drying to 110 degrees Celcius and heating to 550 degrees Celcius. The weight loss on ignition is calculated from the measured weights obtained before and after exposure to the higher temperature. The LOI can be used as a rough estimate of the organic content of the foulant material tested. The results of the LOI indicates that the majority, if not all, of the foulant is organic (biomass).

Figure 1. Example of Membrane Forensics Loss on Ignition Results.

11% Ash (Inorganic material)

89% Volatile (Organic material)

One of the primary causes of membrane failure is surface fouling. Identifying the material fouling the surface is critical for membrane cleaning and/or the development of preventative measures in the system’s pretreatment. Typically, membrane foulants are analyzed by using the following techniques: Fourier Transform Infrared Spectroscopy (FTIR) to identify organic compounds; Scanning Electron Microscopy (SEM) to determine foulant elemental composition; and, Targeted Energy Dispersive X-Ray Analysis (T-EDXA) for inorganic delineation; The autopsy can reveal if the majority of the foulant was organic, most likely living biomass, with typically the remainder being primarily comprised of inorganic matter consisting of calcium, sulfur and phosphorous. If the results of the autopsy show that primarily inorganic material is present, then biofouling may not be a contributor to the fouling process. Autopsies are important to help indicate if a biofouling situation exists within the element. TROUBLESHOOTING OVERVIEW Table 2 illustrates some common troubleshooting problems and presents both symptoms and possible resolutions. Most common of problems in a membrane facility are permeate water quality is poor and/or permeate flow is low (flux has decreased). Both of these situations can be addressed by first evaluating the pretreatment system of the membrane facility. If the permeate quality is poor then profiling (checking) each pressure vessel (PV) is recommended to determine which PV is the problem. Once the suspect vessel has been determined, check the end cap on the downstream side and examine the product o-ring (if it still is in place). If the o-ring is still in place, but flat or damaged, then it should be replaced. If the o-ring is functional, then check the upstream end-cap to see if it still functional. If not, remove and replace o-ring; however, if the o-ring is functional then perform an internal probe of the PV in question should be performed to see which element(s) are having problems.

Table 2. Common Troubleshooting Scenarios and Possible Problem Identification Symptom

Possible Problem

Sudden ↑ in ΔP, no change in rejection

Clogged cartridge filter, blocked pipes or front end membrane element

Sudden ↓ in ΔP, ↓ in rejection

O-ring or brine seal failure, cracked permeate tube

Gradual ↓ in NPF 1st-stage, slight ↑ rejection

Biological or particulate fouling

Gradual ↓ NPF 2nd-stage, ↓ 2nd-stage rejection

Scaling

Gradual ↑ NPF 1st-stage, ↓ 1st-stage rejection

Lead end degradation caused by reaction of Cl2 with transition metals, or advanced biofouling

Gradual ↑ NPF 2nd-stage, ↓ 2nd-stage rejection Advanced scaling Note that differential pressure can also be a cause of low quality permeate water. Differential pressure is measured across the membrane from the feed end to the concentrate end. An increase in differential pressure can indicate that the vexar, the concentrate material that serves as the spacer between the leaves of membranes allowing the water to flow across the membrane. Bacteria tends to grow and scale tends to deposit around this material, which can cause a rise in differential pressure. When the membrane has lost flow, one first must analyze the data to attempt to understand when the membrane started to lose permeate flow. If the temperature of the water drops (cools down), the membranes will start to show a loss of production. The remedy for this condition is to ramp up the VFD slowly and increase the feed pressure of the RO feed pump, adjusting the concentrate flow rate to its normal flow capability. BIOFOULING COUNTERMEASURES The primary means of controlling fouling by mechanism and unit operation are shown in Table 3 and described below: 1. Plugging control is typically required for all RO/NF membrane systems in either surface or groundwaters and is achieved by feed water turbidities and SDI’s less than 0.2 NTU and 2, respectively; 2. Scaling control is typically required for all RO/NF membrane systems in either surface or groundwaters and is achieved by acid and/or antiscalent addition; 3. Bio-fouling control is typically required for aerobic surface or groundwaters and is achieved by NH2Cl or addition of other bactericidal agents; 4. Organic fouling can occur in surface water systems with TOC > 3-6 mg/L and is typically reduced by coagulation, sedimentation and filtration. However, the significance of organic fouling is not often known or predictable.

Table 3. Fouling Control by Pretreatment System Mechanism

Process A/AS Scaling + Plugging Adsorption Bio-fouling* -

MF/UF + + + (feed)

CSF + + -

NH4Cl +

AOC Removal +

*MF/UF will remove biofoulants from feedwater; however, note that MF/UF will not stop growth from occurring within the RO equipment. SUMMARY Productivity is essential to any water treatment facility. Productivity is affected by design of the membrane process and fouling. Designers can select membranes for specific treatment characteristics. Once selected, a designer can select operating conditions for that membrane process. A primary consideration affecting productivity is fouling. The four primary mechanisms of fouling are scaling, plugging, adsorption and biological growth, requiring a continuous, easily adaptable change in unit operations to control fouling. Knowledge of the type of fouling can be used to assist in decision-making on cleaning, restoration and membrane replacement. A complete and thorough membrane element autopsy analysis performed by experienced membrane analytical providers can provide engineers and decision-makers with data and information to determine a solution to operating problems associated with membrane productivity loss. Maintaining a membrane facility requires that detecting problems early can save money and staff frustration. With proper pretreatment, monitoring, troubleshooting, and cleaning of the membrane process, a desalting facility could operate trouble-free for many years and save thousand of dollars in the process. REFERENCES 1. Duranceau, S.J. (2004), Membranes and Biofouling: A Slimy Situation. Proceedings of the 2004 Florida Water Resources Conference, Kissimmee, FL. 2. Duranceau, S.J. (2001), Membrane Practices for Water Treatment. AWWA Trends in Water Series. Denver, CO: AWWA. 3. Beckman, J.E. and M.R. Griffin (2005), “Membrane Forensics: Solving and Preventing Membrane-Related Problems,” UltraPure Water, April, 2005. 4. Pontie, M. et. al. (2005), “Tools for Membrane Autopsies and Antifouling Strategies in Seawater Feeds: A Review,” Desalination, Vol. 181, pp. 75-90.