PFAS, PFOS, PFOA, PVDF?

Debunking Misleading Connections Between PFAS and PVDF Membranes

There is confusion in parts of the U.S. water industry, possibly fueled by non-PVDF membrane suppliers, around the connection of PVDF membranes and new regulations for PFAS compounds in drinking water. PVDF (Polyvinylidene fluoride) is the most widely used material in the manufacture of MF and UF membranes.

The confusion stems from proposed regulations in Europe that would broadly define PFAS as any substance that contains at least one fully fluorinated methyl (CF3) or methylene (CF2) carbon atom (without any H/Cl/Br/I attached to it). This definition would cover a wide variety of chemical structures, including PVDF. In contrast, the current USEPA definitions of PFAS exclude PVDF including the most recent update of the structural definition of PFAS.

The American Membrane Technology Association (AMTA) has stated that PVDF is considered to be part of a class of high molecular weight fluoropolymers that are distinct from non-polymeric PFAS and have distinctly different physiochemical, toxicological and environmental characteristics. AMTA also states that suppliers of PVDF used for water treatment membranes have certified that there is no use of PFAS as processing aids. See the full AMTA Fact Sheet here.

So, with all of this regulatory discussion there is an ‘opportunity’ to confuse those that are not conversant in polymer chemistry and the correct definitions. I have heard concerns from some end users and engineers that PVDF membranes may release PFAS compounds and may therefore be banned in the future. Coincidentally, these rumors are strongest where ceramic membranes are being considered… Ironically, when offered PES membranes as a non-PVDF polymeric membrane alternative, one engineer said they did not want to consider ‘unproven’ membranes, implying that ceramic membranes were considered more proven... I’d say there are orders of magnitude more PES membrane capacity installed than ceramic membranes, also with a lot longer operating history.

If an engineer and owner want to install ceramic membranes, that is their decision to make. I just want to make sure PVDF membranes are not misrepresented and decisions and made based on facts. Firstly, the European regulations are not being proposed due to any concern that PVDF membranes are releasing PFAS into drinking water. They are just a broad regulation to ban production of products using PFAS materials (like Teflon). Secondly, there is no evidence that PFAS compounds are released into drinking water from PVDF membranes after use for over 20 years. As part of the development of regulations for PFAS in drinking water, the USEPA has required extensive testing of drinking water supplies and no connection with systems using PVDF membranes has been made. Thirdly, the USEPA has indicated it has no intention of banning PVDF or classifying it is a PFAS compound, irrespective of what happens in Europe.

I know some water systems will still lean towards using ceramic membranes due to concerns about future regulations against PVDF even though this is highly unlikely. If they are prepared to pay a high premium for membranes as insurance against this low probability, sobeit. Maybe less expensive insurance would be to install an open platform/universal polymeric membrane system now, that can be converted to ceramic if needed in the future.

The comments and opinions in this post are my own and not those of my employer.

What the Flux! – Part 2

Polymeric vs Ceramic Membranes – Lifecycle Cost Comparison Myths

 

In my last post I discussed that just because ceramic membranes can operate at much higher fluxes than polymeric membranes, a ceramic system will not necessarily have a smaller footprint since ceramic modules or stacks have a lot less surface area than polymeric modules – often less than a quarter the surface area in the same footprint. Therefore, higher flux does not necessarily translate into a lower capital cost.

In this post I will focus on the operational cost comparisons and dispel some myths that ceramic membrane systems have lower lifecycle costs due to longer membrane longevity. Let’s start with the longevity. In polymeric versus ceramic membrane lifecycle cost comparisons, membrane replacement frequencies are typically based on warranty lengths which are often 10 years for polymeric and 20 years for ceramic membranes. A 10-year life for polymeric membranes is OK, but there are plenty of examples of Pall (Aria Filtra) and Toray installations still operating with the original modules after 13-15 years. To my knowledge, only Metawater has ceramic membrane installations over 20 years’ old and therefore can guarantee this lifespan. They say the strength of a chain is based on the weakest link and while I don’t doubt the ceramic membranes used in Nanostone, Ovivo and Cerafiltec systems are very durable and would probably last 20 years, what about the plastic components, polymers and glues used to construct and house these membranes? Will these parts last 20 years? I don’t know if any of these systems have been operating more than 3-5 years so far, so it is taking a great leap of faith in believing these systems will last 20 or more years. The housings on the Metawater modules on the other hand are made of stainless steel, so as long as you don’t drop a module (where it would be like dropping a ceramic pot) there isn’t a weak link to fail before the membranes fail.

 The lifecycle calculations I have seen in bid documents are typically over 20 years, and assume a polymeric replacement after 10 years, and no ceramic replacement over that period, so that is where ceramic systems have an advantage. What if we run the lifecycle cost over 20 years and one day, so we include 2 polymeric replacements and 1 ceramic replacement in the lifecycle cost? That’s fair isn’t it?

 From what I have seen, the cost of a ceramic module is at least twice that of a polymeric module while having around a fourth of the surface area. For a replacement cost comparison, let’s assume the following:

• A ceramic module has 1/4 the surface area of a Toray or Dupont polymeric module (see previous post).
• A ceramic module can get 3 times the flux of a polymeric module.
• A ceramic module costs twice as much as a Toray or Dupont module. Say the polymeric module replacement cost is $2600, Ceramic is $5200.
• Let’s say the base case is a 2 x 2 MGD train polymeric system, design flux of 50 gfd, 45 modules per train. Equivalent capacity ceramic skids will have 60 modules (3 x flux but 1/4 surface area per module).
• Cost to replace all polymeric modules is $234,000.
• Cost to replace all ceramic modules is $624,000.

 Based on the above assumptions, for a 10-year polymeric replacement period and 20-year ceramic replacement period, over 20 years and one day, the replacement cost for polymeric membranes is $156,000 less (two replacements of polymeric versus one ceramic replacement). Maybe ceramics can get a little higher flux and maybe the cost assumptions are a little off, but there is no way you are saving much if anything on membrane replacement costs with ceramics. If ceramic membranes can get four times the flux, the replacement cost is just breakeven using my pricing assumptions.

Known Fact: we know there are Pall and Toray polymeric membrane systems that have had their membranes last 13-15 years. Can we be certain the new ceramic membranes on the block will last 26-30 years to break even with the polymeric membrane replacement cost?

Every water is different, and the flux difference may vary one way or the other to shift the economics, but you can’t say there will be a significant savings in membrane replacement costs using ceramics unless you are looking at the polymeric membranes of yesteryear. Kind of like comparing the fuel efficiency of a car from the 70s with today’s modern cars.

I was going to include a comparison of lifecycle costs with coagulant included, where ceramic membranes often need a coagulant dose to achieve high fluxes, while polymeric membranes do not. This can make the lifecycle cost of a polymeric system significantly lower than a ceramic, but I have rambled on for too long already so will leave that to Part 3.

The comments and opinions in this post are my own and not those of my employer.

What the Flux!

Polymeric vs Ceramic Membranes

My sales rep in the Southeast came to me recently concerned that ceramic membrane companies are promoting fluxes to engineers and water utilities in the region in excess of 200 gfd, asking me how polymeric membranes can compete? My response was WTF! Flux is just a number and just because the flux may seem a lot higher than polymeric membranes it does not mean a ceramic system has a smaller footprint or lower cost. I’ve seen a lot of presentations from ceramic membrane companies trumpeting all the reasons ceramic is better than polymeric but I haven’t yet seen a counter argument from a polymeric membrane company or system supplier. So maybe this is the first counter to some of the ceramic membrane company’s claims. I know I’ll get a push back from the ceramic membrane companies because they are all trying to get established in the market, but I can’t just sit back and let the latest polymeric membranes be unjustly grouped with systems of the past. Note that I am not criticizing the integrity or performance of ceramic membranes at all, and there are situations where these are a great fit, but rather I am just trying to provide a balanced and up to date comparison with polymeric membranes.

WTF!*

Let’s start with the high flux claims. I have seen papers on ceramic pilot studies where fluxes up to 200 gfd have been tested but I don’t yet know of a full-scale system in the US that has been put in service with a design flux this high. The largest ceramic membrane system in the US at Butte MT has a design flux of 69 gfd. A ceramic system that was awarded at Mandaree ND a few years ago had a design flux of 120 gfd for summer. These are pressure ceramic systems where there are feed pumps supplying pressurized modules. 

Figure 1: Membrane System Configurations

The other ceramic configuration is submerged flat sheet operating in the vacuum configuration where a pump draws through the membranes (see Fig 1). Companies such as Cerafiltec and Ovivo are providing this submerged technology and are particularly active in the Southeast. I don’t think a flat sheet submerged ceramic system is installed in the US on a full-scale drinking water system yet, but I have seen several pilot study papers. What strikes me about these recent pilot studies is they are not that impressive. I won’t call out any specific studies, but go search the proceedings from recent AMTA/AWWA Membrane Technology Conferences and you will find them (there may be better pilot studies but I can't find any published). They all spend a lot of time optimizing coagulation ahead of the membranes to reduce rapid TMP buildup and then ramp up the flux in steps over 1 to 2 week periods to get to 200 gfd.  I haven’t seen more than a few weeks operation at anything near 200 gfd in the published papers. Whenever I’ve been involved in a pilot study with polymeric membranes it has been necessary to run at stable operating conditions for at least 30 days. Why is the bar lowered when ceramic membranes get evaluated? Note the Mandaree ND ceramic pilot study did have stable operating periods of at least 30 days at 120 gfd.

 Another criteria for setting design conditions for polymeric membranes is to be a little conservative on the design flux compared to what the manufacturers or pilot studies claim is possible, so if a pilot study shows a flux of 60 gfd is possible based on the feed water quality, the engineer will allow 50 gfd for the full-scale system. While the ceramic membrane companies may claim 200 gfd is possible, when it comes to the design, I haven’t seen more than 120 gfd allowed. Even so, polymeric membranes can be disadvantaged from years of full-scale experience and require a more conservative design flux while ceramic membranes can claim aggressive fluxes without past full-scale experience to suggest otherwise.

 Does Ceramic have a Smaller Footprint?

The claim is often made or implied that due to higher fluxes, ceramic membrane systems have a lot smaller footprint. I will prove to you that is absolute baloney! Let’s look at the comparative footprints of polymeric and ceramic systems. For polymeric, I’m going to use Toray’s HFUG-2020AN module which has 969 sq.ft. of surface area and is probably the most popular polymeric membrane on the market currently. Compare this with a Nanostone ceramic module at 258 sq.ft. per module. A Nanostone module has around the same diameter as a Toray module and is around 9 inches shorter, so the footprint of a membrane rack is the same for both modules (ie. a rack with 40 Toray modules is the same size as a rack with 40 Nanostone modules). Therefore, a Nanostone module needs to have 3.8 times the flux of a Toray module just to have the same footprint based on surface area per module. So, if the polymeric module is designed for a 50 gfd flux, the flux through the Nanostone module needs to be 190 gfd to match the Toray module footprint. If the design flux for ceramic is 150gfd, the Toray system at 50 gfd will have a smaller footprint.

 Now let’s look at a Cerafiltec flat sheet submerged system. These membranes are supplied as 64.6 sq.ft. modules that have a footprint of 28” x 22.7”. Based on Cerafiltec’s website, these modules can be stacked in towers 16 high, so that would add up to a surface area of 1034 sq.ft. From the photos I have seen on the website, the tallest I saw was 8 high, but I will be conservative and compare the footprint of a 16 high tower versus Toray modules in the same footprint. The Toray modules are 8.5” diameter, so within the footprint of the ceramic flat sheet tower, you could conservatively fit 4.5 Toray modules allowing for spacing between the modules (see Fig 2). Therefore a Cerafiltec tower at maximum height needs to have 4.2 times the flux of a Toray module to have the same footprint, i.e. if the Toray module flux is 50 gfd, the submerged ceramic system needs to flux of 210 gfd to match the footprint.

Figure 2: Polymeric versus Ceramic Footprint Comparison

So, I hope I have made it clear that flux is just a number and a high flux does not mean that a membrane system will have a smaller footprint. You also need to consider the amount of membrane surface area that will fit within a given footprint and that ceramic modules have a lot lower surface area than polymeric modules. The price of a ceramic membranes compared to polymeric modules on a membrane surface area basis is also a lot higher, so you can’t assume a higher flux also means a lower cost.

 Of course, there are other important considerations when comparing polymeric to ceramic membranes such membrane longevity and lifecycle cost. I have mentioned in a previous post that the longevity of the newer polymeric membranes is much improved over earlier polymeric membranes which has narrowed the lifecycle benefits of ceramic over polymeric. I’ve also calculated that when you consider the requirement of a coagulant dose ahead of ceramic membranes, the lifecycle cost of polymeric membranes can be lower than ceramic. I will elaborate on that in a future post.

*Shoutout to Stuart Leak from Avista who first used this acronym in his presentation What the Foulant.

The comments and opinions in this post are my own and not those of my employer.