By James A. Hunt

Summary: This rinse water reclaim case study contains several valuable lessons. Membranes play a starring role in rinse reclaim, system failures are more often the fault of the system design and, when properly executed, rinse reclaim proves that conservation pays.

This rinse water system was designed at the request of Fosta-Tek Optics Inc., a specialty lens manufacturer in Leominster, Mass. Its production volume has been growing steadily for the past few years and jumped dramatically after 9/11 with a rush of orders for riot shields, gas mask lenses and the like. The once-through, heated deionization (DI) rinse water system was becoming burdensome. The company had attempted a recycle system, conceived and executed by the staff, with disastrous results. The system produced “slime” that left the lenses dirty, killed the DI tanks, and caused frequent shutdowns to dump the system water and clean the wash tanks. Other components of the system failed on a regular basis resulting in high maintenance levels.

The system process flow was filtered city water (5 micron), DI exchange tanks, heater, split flow to a batch-filled sonic cleaner, a first-rinse tank, and a two-compartment cascading second and third-rinse tank, then to the city sewer. The data summary is shown in Table 1.

All systems go
The first step in system design is to determine what it is, and establish what it should be. This can’t be done in the office but requires extensive interviews with plant engineers and floor personnel that will use the system. Fosta-Tek was happy with the quality of the once-through DI system that was in place but wanted a higher flow rate and lower operating costs.

The limitation to flow was the operating capacity of the inline heater. The heater was producing an 85°F rise or a net output of 4,248 British thermal units per minute (BTUs/min). It stands to reason that recycled water at 105°F mixed with makeup water at 50°F will result in warmer water and the possibility of running a higher flow rate through the heater. A quick calculation shows that a 24 percent increase in feed water (recycle plus makeup) temperature will make it possible for the heater to keep up with demand.

Recycling the idea
The second step in the design process was to determine the impact of recycling product water quality. We bench-tested the rinse wastewater. The first step was to run the water through a mixed bed and measure the total dissolved solids (TDS) level. We found that ion exchange reduced TDS by 75 percent resulting in water at, or very near, the 1-megohm product quality level; however, in a recirculating system, the TDS creep would soon have us over the limit. A low-pressure reverse osmosis (RO) membrane removed >95 percent of the TDS. We deduced from these tests that not all of the contaminants were ionically charged. We believed the major contaminant to be the surfactant and, although it was “anionic,” there was probably a significant organic component. The surfactant manufacturer declined to share the product’s formula but we had no reason to suspect monovalant ions. Guided by the success of the low-pressure (60 pounds per square inch, psi) RO membrane tests, we selected a nanofiltration (NF) membrane. The salient considerations were the high rejection (>95 percent) of divalent ions, organics and other weakly charged contaminants, the high recovery rate (75 percent) and low operating pressure (<150 psi). This was sized for 8 gallons per minute (gpm) permeate as follows:

  • 6, 4 x 40 spiral-wound membranes
  • Feed flow = 10.6 gallons per minute (gpm)
  • Feed TDS = 66.2 parts per million (ppm)
  • Recycle flow = 9.4 gpm
  • Concentrate flow = 2.6 gpm
  • Concentrate TDS = 245.5 (ppm)
  • Product flow = 8.0 gpm
  • Product TDS = 7.6 ppm

Mixed-bed exchange tanks—combining strong acid cation and strong base anion (SAC/SBA) resins— then followed the nanomembranes. These consist of two 14-inch, 3.6 cubic foot tanks with two in each series.

The third step in the system design was to address the operating costs. These costs are comprised of a short but critical list—water and sewer, heating (kilowatts per hour), DI exchange tank service, maintenance, production downtime, and environmental and public relations.

As in most places, the water and sewer costs are significant. In this study, at the desired flow rate of 8 gpm, water and sewer would cost $65.70 per day in a once-through system. Heating is an even greater burden at $300.52 per day. DI exchange tanks were the monster at $765 daily. Maintenance costs weren’t tracked but consisted of changing cartridge filters every other day and changing out DI tanks every four days. Production downtime costs weren’t shared but were considered the highest operating cost. Leominster, when this was written, had a mandatory water ban in effect. With or without economic incentives, the management of Fosta-Tek was predisposed to conserving water.

Slashing the costs
Almost everything points to recycling as the solution to the high costs of the current system. Recycling reduces water, sewer and heating costs. Even DI exchanges are curtailed, as the rinse wastewater is much lower in TDS than the city water supply (15 ppm vs. 250 ppm). Recycling reduces the total water demand, which allows Fosta-Tek to do its part to ease the water shortage. The real design considerations with regard to cost revolve around production downtime and maintenance. Certainly, Fosta-Tek’s own experience demonstrated how quickly these items could scuttle a project.

Key to a successful design is the elimination (reduction) of biogrowth. We took a two-step approach. We included in-line ultraviolet (UV) lights to keep biogrowth down. The second step was an automatic sanitizing system. This was proposed to limit the need for plant personnel to sanitize the system between the hours of 4 a.m. and 6 a.m. during the week and weekends. The automated system controlled by the system microprocessor is very user friendly. At the push of a button, motorized valves open and close to bypass chlorine intolerant equipment, specifically the nanomembranes and the DI exchange tanks. Chlorine is then injected into the circulation water stream. Total chlorine at 100 ppm is circulated throughout the system including the rinse tanks for a selected amount of time (one hour). At the end of the chlorination cycle, valves change position diverting the flow through a carbon tank to dechlorinate the water. After the dechlorination cycle, valves again change position to put the system back in the normal operating mode. Other than initiation, no operator assistance is required.

Other maintenance items include cartridge filter changes, which happen a third as often because the makeup water has been reduced by 75 percent. The vendor changes the DI tanks once a month instead of the wash line operator changing them every few days and the vendor, on a service contract, does UV lamp replacement and membrane cleaning.

Pairing them off
Production downtime due to maintenance requirements and/or equipment failure cannot be tolerated. To avoid this contingency, every major component has been duplicated; that is, every pump, UV light, filter housing, and NF unit is installed in pairs. In normal mode, the components alternate operation each time the system is turned on (daily). In the event of service or repair, the system is changed from automatic to manual and the component is selected electronically to run. In addition to duplicating components, if two components fail at the same time, a bypass exists allowing the system to be put back into a once-through DI exchange tank mode. Such precautions are because, when it comes to components with moving parts, it’s not a question of if they break down but when they break down.

Operational cost savings are significant in most rinse recycle situations and particularly in this case. The savings here are computed as though the current system operated at 8 gpm. The savings don’t take into account maintenance labor savings since there’s no record of previous maintenance costs. The comparison is shown in Table 2.

The savings are $907.59 per day, $4,537.95 per week, $19,513.19 per month and $234,158.22 per year. The system costs $80,000, resulting in a payback in 4.1 months.

Our experience with this rinse water reclaim application confirms several beliefs. Membranes play a key role in recycling water and NF is particularly well suited to this situation. All water reuse systems need frequent sanitization and an automated one can be more reliable than operator assisted. Membranes seldom fail in a properly sized and designed system; however, mechanical components do fail—so redundancy and/or operational options need to be included.

This article is based on a paper presented at the 20th Membrane/Separations Technology Planning Conference, November 2002, Newton, Mass. The author thanks event organizer BCC Inc. for permission to reprint it here.

About the author
James A. Hunt is president of DI Water Inc. and principal manager of The Water Group LLC, both of Amesbury, Mass. The company partners with water treatment dealers in the design of systems as described here. He’s also a past member of the WC&P Technical Review Committee. He can be reached at (800) 840-0901, (978) 834-3169 (fax), email: or website:


Comments are closed.