Virus and faecal coliform removal

For disinfected tertiary effluent, California Title 22 regulations require a minimum chlorine CT value of 450mg-min/L or a 5-log virus inactivation in addition to the effluent total coliform concentration not exceeding a 7-day median of 2.2 most probable number (MPN)/100 mL (Hirani, 2010)

Since majority of naturally occurring viruses exist as an embedded form in microbial floc structures, virus removal efficiency is generally high regardless of membrane pore size. In addition, dynamic membrane plays a great role in removing virus. The initially formed cake layer on membrane surface can reject the viruses smaller than membrane pore sizes. Therefore, virus removal efficiency is much less affected by membrane pore sizes than the relation between the sizes of virus and membrane pore suggests. Following are some experimental observations found in literature.

  • When phage T4 (0.11 μm) was used as a model particle, a membrane with a pore size of 0.22 μm resulted 5.7 log removal initially, but it increased to 8.0 log removal as cake layer formed (Lv, 2006).
  • The removal efficiency of naturally occurring MS-2 Coliphase could be only weakly correlated with membrane pore sizes as shown in Fig. 1 and Table 1 (Hirani, 2010). The weak correlation can be attributed to the dynamic membrane effect and the embedded viruses in flocs.
  • When MS-2 coliphage (0.03 μm) was spiked to an MBR treating municipal wastewater, pore size effect became more apparent  (Hirani, 2010). The microfiltration membrane with 0.1 μm pore size showed 98.0% removal (or 1.7 log), but the membranes with 0.03 μm pore size showed 99.996% removal efficiency (or 4.4 log). Perhaps the spiked foreign viruses did not have enough time to be embedded in microbial floc until they are filtered by membrane. The improvement of removal efficiency with tighter membrane can be critical for sanitary purposes, but it is mere 2% increase in terms of removal efficiency.

Virus 18Fig. 1. Indigenous MS-2 coliphase removal by the MBR systems evaluated in the study. Nominal pore sizes are A-0.05 μm, B-0.04 μm, C-0.08 μm, D-0.03 μm, E-0.1 μm, F-0.1 μm (Hirani, 2010)

As bacteria, faecal coliforms are much larger than viruses and can be filtered by 0.45 μm filters. Therefore the data summarized in Table 1 does not show any trend between faecal coliform removal efficiency and pore size (DeCalolis, 2007). It was pointed out that microbial colonies in the internal space of permeate pipe line were the major source of the microorganisms detected in permeate. Effluent turbidity and COD did not show any strong correlation with pore size since particle removal efficiency is hardly affected by pore size due to  dynamic membrane (or cake layer). In other study (Nishimori, 2010), the trend of high coliform removal efficiency maintained for more than 10 years in two full-scale municipal MBR plants, where 5-log removal of coliform was achieved with 0.4 μm flat sheet membranes.

Table 1. Effect of pore size on permeate quality (Reproduced from DeCarolis, 2007)

 UnitsKubotaUS FilterZenonMitsubishi
Commercial designation Type 510MemJet B10RZW 500 DSterapore HF
Configuration Flat SheetHollow FiberHollow FiberHollow Fiber
Material PEPVDFPVDFPE
Pore size 0.40.080.040.4
Effluent Turbidity*NTU0.08 ± 0.020.04 ± 0.020.06 ± 0.020.07 ± 0.02
Effluent COD*mg/L18.4 ± 9.620.5 ± 13.417.3 ± 8.623.2 ± 5.3
Effluent Total Coliform*MPN/100mL13 ± 69386 ± 674807 ± 13147 ± 7
Effluent Total Coliphage*PFU/100mL10 ± 2413 ± 131 ± 113 ± 22

* Average ± Standard Deviation

The virus rejection efficiency varies depending on the age of membrane perhaps due to reversible and irreversible membrane fouling in membrane pores and surfaces (Tazi-Pain, 2006). An example is shown in Fig. 2, where 2-month old membranes have higher MS2 phase removal rate than new membranes. It is also noticeable that the initial removal rate is somewhat better than that in later stage of filtration, but the exact cause is not known.

Virus 8
Fig. 2. Virus removal efficiency of new and old membranes at different time frames in a filtration cycle (Tazi-Pain, 2006).