The concept of FO has been known for thousands years, but it was only in the middle of 20th century that FO was suggested as a water purification method. But, the research on FO as a water purification process had been virtually dormant until recent decades due to the lack of suitable membranes. In the meantime, reverse osmosis (RO) has had a remarkable success since the invention of commercially viable membranes in 1960s followed by the invention of spiral wound modules in 1970s. As RO technology has been matured, FO has been drawing rejuvenated attentions as a future desalination solution since 2000s.
In FO, permeate is drawn through FO membrane by using a draw solution that contains concentrated salts, e.g. NaCl, MgCl2, MgSO4 etc. The diluted draw solution can be concentrated again by RO, where permeate becomes a final product and concentrate becomes a rejuvenated draw solution. Alternatively, volatile salts such as ammonium carbonate can be used as salt. In this process, ammonium carbonate is recovered by heating the diluted draw solution. Substantial level of residual ammonium carbonate remains in the product water in practical conditions unless multistage distillation or stripping processes are employed.
Despite of the efforts put in during the last decade, FO is far from commercial success due to the following reasons. More details are found here.
1. Unfavorable thermodynamics
FO has often been misunderstood as a more energy efficient process than RO, but it is thermodynamically not possible. There is no practical way to make FO less energy intensive than RO unless draw solution recovery is not necessary. .
Let’s take a look at the FO process combined with draw solution recovery process from the perspective of thermodynamics. In FO process, free energy decreases first in FO because it is a spontaneous reaction. Then, the diluted draw solution must be concentrated to recover draw solution from the product water, where free energy of the diluted draw solution must be raised. As illustrated in Fig. 1, extra free energy is necessary in draw solution recovery process due to the initial decline in FO.
In contrast, free energy of water directly increases when water is purified in RO and hence much less energy is necessary for the process. As a result, there is no way FO is more energy efficient than straight RO.
Though the energy demand is high, “energy costs” can be less for FO, if volatile salts are used as draw solution and waste heat is used to recover the salts. However, recovering the salts thoroughly from the product water is not possible without putting mutistage recovery processes. Then, the initial purpose of realizing cost effective desalination is defeated by itself.
Figure 1. Conceptual comparision of free energy changes of feed water during the FO and direct desalinations, where external energy is required to compensate the free energy gains by water (Yoon 2015).
2. Detrimental loss of driving force caused by concentration polarization
In any membrane processes, separation occurs on the membrane skin that contacts with feed water. In order to minimize the filtration resistance, the skin layer must be kept thin as much as possible. FO and RO are not an exception. The skin layers are cast on porous support layers that consist of intermediate layer and support fabric.
In RO, feed water contacts only the skin layer and the permeate fills in porous support layers, thereby concentration polarization (CP) occurs only in feed side of the membrane.
In FO, however, concentration polarization occurs in both side of the membrane. If skin layer is facing feed water, CP occurs in feed just like in RO. In addition, CP also occurs in the porous support layer, where fresh water meet with salt water. Since the turbulence in the bulk stream does not effectively relieve the CP in the support layer, CP causes a detrimental loss of driving force. On the contrary, if support layer faces feed water, foulants will plug up the pores of support layer causing a serious filtration resistance. Meanwhile, the permeate emerging on the skin layer dilutes draw solution and diminishes the driving force.
The loss of driving force caused by CP is detrimental with respect to the energy efficiency in FO. For example, even if 0.5M NaCl solution is used as draw solution, which is equivalent to ~ 25 bar of transmembrane pressure, flux is typically no greater than 10 LMH. This starkly contrasts with RO, where flux is typically 20-40 LMH at 10 bar. If permeabilities are compared, <0.4 LMH/bar and 2-4 LMH/bar for FO and RO, respectively.
3. Reverse salt diffusion
In water continuous phase, salts permeate through membrane by diffusion. The amount of salt passing the membrane is functions of salt concentration and contact time. Due to the high salt concentration in the draw solution, a large quantity of salt diffuses back to feed water. Since feed water must contact with membrane longer in FO than in RO due to the low flux, the amount of salt reverse diffused to feed becomes even more. Naturally, a large quantity of salts is lost from draw solution. If FO membranes are applied to MBR, where draw solution is circulated through FO membranes submerged in aeration tank, a large amount of salts will diffuse to biological system and damage the biological stability.
4. Excessive capital costs
There are perceptions that FO is less susceptible with membrane fouling, but the low fouling is just a natural consequence of the low flux. As discussed here, here, and here, particulates in feed water tend not to deposit on membrane surface at low flux due to the critical flux caused by the back transport phenomena. Even if some particulates deposit, they tend not to be compacted too much at low flux and would not quickly cause filtration resistances. As a consequence, membranes appear not to foul easily.
A drawback of the low operating flux is a larger membrane surface areas necessary to treat the same feed flow. If the flux is 10 LMH in FO, 100% more membrane surface areas are necessary comparing to the RO running at 20 LMH. If the capital costs for draw solution recovery are considered, overall capital costs will increase even further.
© Seong Hoon Yoon