The Voice of the Finishing Industry since 1936

  • PF Youtube
  • PF Facebook
  • PF Twitter
  • PF LinkedIn
10/1/1997 | 14 MINUTE READ

Aqueous Cleaner Filtration Study

Facebook Share Icon LinkedIn Share Icon Twitter Share Icon Share by EMail icon Print Icon

Filtration can help remove contaminants and extend cleaner life...

As chlorofluorocarbons (CFCs) are phased out globally, alternative aqueous cleaning systems are being implemented by the metalworking industry. These aqueous systems require significantly larger volumes of cleaning and rinsing solutions than did the solvent systems they replaced. In an effort to minimize chemical waste, fluid management is suggested.

There are several steps involved in fluid management. These include selecting an aqueous cleaner that rejects rather than emulsifies soils, filtering to remove contaminants and routine product additions to insure that an adequate amount of cleaner is present in the cleaning bath.

This study focuses on the filtration portion of the overall management program. A major aerospace subcontractor selected a high-pH, multi-metal-safe, alkaline cleaner for use in an agitated immersion tank. This cleaner was used at a five pct dilution with water. The soiled or contaminated cleaning bath was then filtered through five different types of media supplied by four vendors.

Feed, retentate and permeate samples were analyzed to determine the chemical and performance properties of each compared to a virgin sample. This study depicts the influence filtration equipment, media and design type and pore size have on the cleaner properties.

Water volumes problematic. The industrial manufacturing industry has made significant strides in substituting aqueous cleaning processes for banned ozone depleting chemicals (ODCs) and other volatile organic compounds (VOCs) regulated by air quality boards throughout the U. S. These aqueous cleaning processes use greater fluid volumes than were traditionally used by vapor degreasers. Frequently one or more rinses are also needed as part of the new cleaning process, again increasing the total fluid volume.

The increased fluid volumes could be problematic from a disposal point of view. Unnecessary disposal costs could also be prohibitive if continuous or frequent fluid dumping is part of the process. Fluid management and recycling practices are encouraged to minimize waste and reduce disposal costs. Filtration is frequently part of the fluid management process.

Filtration effects studied. A major aerospace subcontractor has converted from hydrocarbon solvent to an aqueous cleaning system. In an effort to prolong bath life and minimize waste, this user selected an aerospace-qualified aqueous cleaner having soil-rejecting properties. For this study, the cleaning bath was used for daily parts cleaning. After a week or so, a quantity of wash solution was then removed for filtration with each experimental filtration system.

Filtration systems and suppliers. Four filtration suppliers provided five filtration processes for this study. The first filtration process was the Silverback1™ from U.S. Filter, an aqueous cleaner recycle system
using a Membralox Pl960 ceramic membrane with a pore size of 0.2 micron. According to the manufacturer2, the dirty fluid is directed through a bag filter to remove coarse particulates. The solution then goes into a settling compartment where the oils and sludge separate out. Drawn from the center of the settling compartment, the dirty fluid then goes to a processing compartment where contaminants are concentrated. Process fluid is pumped at high recirculation rates through channels of the ceramic membrane where the contaminants are separated from the cleaner. The purified cleaner is then returned to the cleaning tank.

The UF-4 pilot filtration unit, from Koch Membrane Systems, uses two different polyvinyl difluoride (PVDF) 100,000 molecular weight cut-off (MWC), through-flow tubular membranes. According to this supplier3, the membrane filters are polymeric coatings or extrusions with inverted conical-shaped pores. Dirty fluid is pumped across the membrane at high flow rates. The parallel fluid flow eliminates the cake-like film that is typical of conventional filters, thus extending the life of the filter.

The third process was the Waste Wizard from Membrex, a vortex flow filtration unit that uses a 50,000 MWC MX-500 flat-plate membrane. The manufacturer describes the system as having a special membrane system for removing contaminants from a cleaner bath without pretreatment using coalescers or skimmers.4 This filtration process also relies on crossflow over a membrane.

The Sanborn Aquamate 21 filtration unit, with a Dell PTFE spiral-wound membrane, has a pore size of 0.2 micron. Product literature from this supplier describes the process as liquid pressure driven across a semi-permeable membrane in a crossflow pattern.5 Contaminants are concentrated and then removed from the process tank while the clean permeate is returned to the cleaning tank.

An unfiltered sample was retained from each experiment. Retentates, collected where applicable, and permeates were compared to both the feed sample and fresh cleaner control sample. Although the wash solution is used at an elevated temperature, it is not documented whether these test samples were filtered at ambient or elevated temperature.

Soil-rejecting cleaner chosen. The rejection-type cleaner used in this study is a high-alkaline aqueous blend of sodium, potassium and silicate salts with organic surfactants and inhibitors. Since tap water is used in the cleaning process, calcium and magnesium salts were also monitored. Along with the elemental analyses, additional properties monitored include pH, cleaner concentration, alkalinity ratio, refractive index, conductivity, total solids and suspended solids. The organic components were quantitatively measured and reported as Surfactant A (SA) and Surfactant B (SB).

Standard test methods used. Standard test methods were used throughout this study. ASTM methods were used when appropriate. Conductivity and pH meters were calibrated according to methods specified by the manufacturer, as was the Refractometer, which measured degrees fluoride flux (BRIX).

Cleaner concentration was determined using an acid-base titration according to methods recommended by the chemical supplier. The alkalinity ratio, which measured free and total alkalinity at two end-points, was also determined using acid-base titration. Total and suspended solids were measured gravimetrically.

Elemental analyses were run using induction-coupled plasma (ICP) atomic fluorescence spectroscopy. Organic content was measured using infrared (IR) spectroscopy. An HORIBA LA10 laser scattering particle size distribution analyzer was used to determine particle size.

Interpretation of data. Each feed solution showed an increase in the ratio of free to total alkalinity. Filtration had minimal impact on the alkalinity. The conductivity of these solutions showed a significant increase versus a fresh lab sample. Again, filtration showed minimal impact on the electrolyte content of these solutions. All other parameters were affected to some degree by the filtration process.

Test results. The alkalinity ratio, which expresses the ratio of free to total alkalinity, showed an increase in all feed samples when compared to a fresh lab standard. This ratio was minimally affected by filtration (Table I).

TABLE I -- Filtration Study: Raw Data on Feed, Retentate, Permeate and Control Samples
Size µ











      Control 10.9 3.2 1.1 0.1 928 4800 10 12.9 1.0 84 198 79
U.S. Filter
  Feed 9.6 3.2 2.3 0.1 1690 8200 2200 2.6 11.9 68 336 99
0.2 Retentate 9.9 3.1 2.5 0.2 1720 9600 3800 3.1 16.00 72 364 106
      Permeate 10.6 3.1 2.4 0.2 1740 6000 1600 2.8 11.4 65 334 102
      Control 10.9 3.1 1.1 0.1 903 4650 10 12.6 1.0 82 193 77
  Feed 10.1 3.1 2.5 0.1 1660 7150 1800 3.9 12.1 66 335 102
10.0 3.2 2.5 0.1 1760 11200 5000 3.4 15.0 62 377 114
10.0 3.1 2.4 0.3 1720 6800 1480 2.3 9.5 65 103 333
9.9 3.0 2.6 0.1 1750 19600 420 3.7 15.7 68 367 111
10.2 3.0 2.4 0.7 1600 6600 1600 2.1 8.5 64 97 318
      Control 10.9 4.4 1.5 0.2 1290 6600 10 18.0 1.0 117 275 110
Waste Wizard
Vortex flow   Feed 10.2 4.4 2.5 0.7 3060 8800 2000 12.6 14.2 14 181 538
flat plate
Permeate 10.2 4.6 2.3 0.6 2820 7600 1600 12.6 13.3 14 178 520
      Control 10.9 4.7 1.6 0.2 1380 7050 10 19.3 1.0 125 294 118
Aquamate 21
Siral wound   Feed 10.1 4.7 2.3 0.6 2450 9600 2000 12.9 14.5 15 184 545
  Dell PTFE 0.2 Retentate 10.0 4.6 2.4 0.7 2520 10200 2800 15.7 13.1 17 184 552
      Permeate 9.9 4.4 2.4 0.7 2480 8200 1200 14.7 12.0 21 179 532

The feed samples of the U.S. Filter and Koch tests showed no change in the refractive index compared to a fresh lab sample. However, both the Membrex and Sanborn tests showed an index of refraction several times that of the control. An increased value was typically seen when solution solids increased. This could be an indicator of increased product solids or contaminants.

As a measure of the electrolyte content of a fluid, the conductivity value will increase as dissolved salts increase. The conductivity can be affected by both the product solids level and contaminants, including water salts and soils. The electrical conducting properties of all feed, retentate and permeate samples were increased versus a fresh lab standard. The conductivity was minimally affected by filtration methods included in this study.

Total solids is the total amount of residue remaining after drying at 105C for four hours, measured gravimetrically. This value represents all soluble and insoluble solids contributed by the cleaner, dilution water and contaminants. Filtration would remove a significant amount of the insoluble material but would have no effect on soluble components.

Suspended solids is the total amount of undissolved material greater than 0.45 micron. Typically, suspended solids increase as contaminants are introduced into the bath. Filtration should significantly reduce the amount of suspended solids present in the fluid.

Calcium is most likely from the water used to dilute the cleaner chemistry in the cleaning bath. All feed, retentate and permeate samples had values less than the laboratory prepared solution. The level was not affected by any of the filtration processes.

All feed samples showed a ten-fold increase in magnesium. This may be due to the water used as make-up at the facility or could come from a soil being introduced into the cleaning bath with use. In each filtration process tested, the level of magnesium present in the retentate was higher than in the permeate, suggesting that some of the magnesium present in the feed stocks was insoluble, perhaps as a magnesium soap.

The cleaner is the primary source of silicon. The feed solutions used in the U.S. Filter and Koch tests showed a slight depletion of silicon when compared to the control. The feed stocks for the Membrex and Sanborn tests showed a significant depletion of silicon. Both retentate and permeates showed little change from the feed sample. Consumption during use had a greater impact on this cleaner component than did filtration.

Again, the cleaner is the primary source of potassium. Potassium had the opposite result when compared to silicon. The feed solutions showed a significant increase in the potassium levels in the U.S. Filter and Koch tests. The Membrex and Sanborn tests showed the feed stock to have had a significant reduction in potassium. The U.S. Filter, Membrex and Sanborn filtration processes had a negligible impact on potassium. However, both Koch membranes tested showed a significant decrease in the potassium level following filtration.

The cleaner also contributes significantly to sodium. All feed solutions showed increased values when compared to a lab standard. This could be due to concentration of product solids, but most likely is due to a combined increase in product solids as well as contaminants. Again, the Membrex and Sanborn tests showed higher values than the U.S. Filter and Koch tests. Only the Koch membranes appeared to affect the sodium in the cleaning bath. The permeate from both membranes showed a threefold increase in sodium.

When quantitatively measured with infrared spectroscopy, there was a slight reduction of Surfactant A with the feed samples used in the U.S. Filter and Koch tests and a moderate reduction with the Membrex and Sanborn systems. With the latter systems, the permeate surfactant level was still higher than the control sample. Therefore, cleaning efficiency should not be impacted. The moderate reduction in the U.S. Filter and Koch samples would be readily made up with routine concentrate additions. No reduction in cleaning efficiency would be expected.

In regards to Surfactant B, feed samples from the U.S. Filter and Koch tests were comparable to the unused laboratory control. The feeds from the Membrex and Sanborn tests were greater than found in the control, suggesting concentration of the surfactant during use or the addition of a soil that has a similar infrared absorbency. All but the Membrex system showed a reduction of Surfactant B during filtration. The net residual was still adequate and would not likely cause a deficiency in overall cleaning performance. Routine concentrate additions should make up for any surfactant lost during filtration.

The amount of particulate contaminants was significantly less in the permeate than in the control solution. Surprisingly, all test solutions, except the control, showed a presence of some particulate matter, even after filtration. When measured by laser scattering distribution, feed stock samples showed particulates ranging from 0.6 micron to 450 microns. The Membrex system showed a slight reduction in the mean and median particle size. The other four filtration processes showed an increase in the average particle size in the permeate. Table II shows the average particle size for each test solution. It also gives the particle size variation found in the test fluid. Figures 1 through 4 graphically illustrate the particle size distribution for each filtration process versus the control solution.

TABLE II -- Filtration Study: Particle Size Distribution
Supplier Media Size µ Sample
Mean Particle
Size, µ
Median Particle
Size, µ
Particle Size
Range, µ
      Control 0.8 0.8 0.1 to 4.0
U.S. Filter Silverback Ceramic membrane   Feed 75.8 69.7 0.9 to 300
  Membralox P19-60 0.2 Retentate 39.3 1.9 0.2 to 300
      Permeate 66.5 59.9 15.0 to 300
      Control 0.8 0.8 0.1 to 4.0
Koch UF-4 Through-flow tubular   Feed 54.6 56.4 0. to 200
  PVDF HFP-276 100 MWC Retentate 276R 30.9 2.2 0.7 to 200
  PUDF HEM-251   Retentate 251R 14.9 1.7 0.2 to 150
  PYOF HFP-276P   Permeate 276P 71.8 64.1 15.0 to 200
  PVDF HFM-251   Permeate 251P 67.0 64.3 15.0 to 175
      Control 0. 0.8 0.1 to 4.0
Membrex Wastw Wizard Vortex flow   Feed 80.9 70.0 0.6 to 450
  MX-500 flat plate 50K MWC Permeate 60.5 56.5 9.0 to 200
      Control 0.8 0.8 0.1 to 4.0
Sanborn Aquamate 21 Spiral wound   Feed 65.0 62.3 0.7 to 200
  Dell PTFE 0.2 Retentate 49.9 49.4 0.6 to 250
      Permeate 67.2 61.8 0.7 to 250

Filtration affects some properties, but not others. An increase in alkalinity ratio is typical as the alkaline cleaner is used. The alkalinity ratio is indicative of cleaner consumption by soils and is minimally affected by filtration. Conductivity also is minimally impacted by filtration. Conductivity is affected by dissolved water salts, product solids and some soils. It can be reduced by dilution, not filtration.

The U.S. Filter filtration process had a negligible effect on the cleaner chemistry. Only a modest reduction of surfactant was evident. Although the amount of precipitate was reduced, the average particle size remained constant.

Both Koch filtration systems showed an increase in the light refracting properties of the permeates. A reduction of all elements, with the exception of sodium, and both surfactants was also seen. Particle size was increased, and the range of particle sizes was reduced.

Due to the design of the Membrex filtration system, no retentate was available. Only the before and after samples were able to be evaluated. There was virtually no change in the elemental composition of the cleaner solution. There was a modest reduction in the surfactant levels and slight reductions in the average size and range of particulate following the filtration process.


The Sanborn unit also had a minimal effect on the elemental composition of the cleaner bath. There was a slight reduction in Surfactant A. However, the measured amount was still greater than the control sample at the same concentration. There was little difference between the feed and permeate particle size or range. A summary of these results is shown in Figure 5, at right.

Filtration extends bath life. Only the Koch membranes reduced the elemental and surfactant levels to a measurable degree. The other three filtration systems did not adversely affect the elemental composition. However, all reduced the surfactant levels somewhat.

All filtration processes reduced the amount of contaminants in the bath. Although there was a measurable reduction in the variation of particle size with filtration, there was little effect on the average size of the particulate in these bath samples.

None of the filtration processes evaluated would negatively impact the overall cleaning performance of the liquid. Each would be beneficial in removing gross contaminants without appreciably removing cleaner components. Traditionally, the elements and surfactants reduced would be sufficiently replaced through routine additions of cleaner concentrate.

Routine removal of contaminants through filtration extends the functional life of a chemical cleaning bath while reducing the amount of liquid waste, without significant adverse effects on the cleaner chemistry, independent of the type of system chosen. 



  1. Silverback™ and Membralox® are registered trade names of U.S. Filter.
  2. United States Filter Corporation, Warrendale, PA. Bulletin No. WAMS-0296.
  3. Koch Membrane Systems Inc., Wilmington, MA. 4-96/5M KPN 0678277.
  4. Membrex® is a registered trademark of Membrex, Inc., Fairfield, NJ., "Alkaline Wash Recycle System."
  5. Sanborn Technologies, A Waterlink Company, Medway, MA., NEP-10/96~5M.