While MRI has traditionally been used in the medical field for non-invasive imaging of humans or animals, engineers and scientists have begun to tap into this valuable technology for non-invasive imaging of transport and structural parameters inside water systems, due to the abundance of H1 nuclei (in H2O).
Research on contaminant transport and fate traditionally relies on sampling at the outlet of the system thus producing breakthrough curves which are used to elucidate transport and removal processes (Pang and Close 1999; Liu et al. 2005;
Hatt et al. 2007; Werth et al. 2010). However, this approach cannot reveal the spatial heterogeneity of structure and transport inside the system which often hold the key to its behaviour. MRI has allowed researchers concerned with water systems such as rivers (Haynes et al. 2009), filter membranes and water
RF pulse
X
Y Z
B0
Mz
Excite mode
X
Y Z
B0
Mzy
ω0 = γB0 Receive mode
___________________________________________________________________________________
101
treatment systems (von der Schulenburg et al. 2008a; von der Schulenburg et al.
2008b; Creber et al. 2010a; Creber et al. 2010b; Pintelon et al. 2010) and currently SuDS to look inside at the processes involved in flow, biofilm growth and contaminant removal in more detail (Werth et al. 2010).
Early research utilizing MRI in engineering science aimed to look at pore structure and distribution within a porous media, which is important in systems such as oil recovery, remediation of soil or mass transport in packed beds (Baldwin et al. 1996). Localized measurement of porosity of rocks using MRI was first reported in Rothwell and Vinegar (1985), Vinegar (1986), Hall and Rajanayagam (1987), Lamrous et al. (1989) and Merrill (1993) with further characterization methods of void space described in Baldwin et al. (1996) and pore structure and connectivity methods refined in Doughty (1998).
Further research with MRI for characterization within porous media then began to look at fluid transport processes within such systems. Detailed flow mapping of fluids in sandstone was reported in Guilfoyle et al. (1992). Sederman et al.
(1997) used MRI imaging and velocity flow measurements to determine structure-flow correlations in a packed bed of ballotini (glass beads), visualizing channelling effects, laminar flow and constant flow depending on the structure of the pore space, concluding that flow is not homogenous within a packed column. Sederman et al. (1998) further advanced this research with enhanced MRI resolution allowing for flow monitoring throughout the column which also demonstrated significant heterogeneity dependent on local pore space geometry and correlated to the local Reynolds number (in reference to flow rate). Nestle et al. (2003) examined the spatial and temporal adsorption of heavy metals within a sand column, successfully visualizing adsorption and remobilization of heavy metal (Cr, Cu and Gd) contaminants. In this guise, further advancing the understanding of heavy metal immobilization processes with MRI, Phoenix and Holmes (2008) investigated Cu transport within a naturally occurring biofilm and was able to demonstrate that MRI could effectively determine biofilm structure, diffusion coefficients, and map Cu concentrations. Cu concentrations were successfully calibrated to spatially describe immobilization of the metal within the biofilm. This enabled the generation of a model which determined the dominant Cu transport and immobilization processes (found to be dominated by diffusion and adsorption as opposed to advection and precipitation). However,
Chapter 4 MRI of biofouling and metal transport in a gravel filter ___________________________________________________________________________________
102
these experiments were undertaken at the small-scale with a 1 cm thick biofilm grown in a 1.5 cm column at 200 àm resolution which is unrealistic for direct application of results to coarse-grained SuDS filters. MRI was also successfully utilized to spatially image distribution of biofilms for detection of biofouling within membrane filtration systems (von der Schulenburg et al. 2008b) and a bioreactor (Seymour et al. 2004a). Here, MRI technology was able to non- invasively quantify effective surface area of the membranes which is key to design and operation of important water treatment in which biofilms can affect mass transport and hydrodynamics of the treatment systems. Yet, again this research was undertaken on a 55 mm filter membrane and 1 mm bioreactor which is unrepresentative of SuDS.
While the specific studies detailed above are directly relevant to the present thesis, it is worth comment of the handful of papers associated to the wider use of MRI for biofilm research (largely motivated by biofouling of medical equipment and thus undertaken at the very small scale of microns through millimeters). For a wider review of biofouling research using MRI in porous media in the medical field papers include: Hoskins et al. (1999); Seymour et al.
(2004b); Seymour et al. (2007); Shamim et al. (2013)). Further, related studies on use of MRI to image biofouling in membranes can be found in Seymour et al.
(2007); Creber et al. (2010a); Creber et al. (2010b); Pintelon et al. (2010), whilst MRI use for metal transport processes within biofilm can be found in Beauregard et al. (2010); Bartacek et al. (2012); Cao et al. (2012); Schulenburg et al. (2008);
Bartacek et al. (2009); Ramanan et al. (2010); Vogt et al. (2012); Ramanan et al.
(2013). Finally, studies looking at the structural form and architecture of biofilms themselves include Manz et al. (2003); Neu et al. (2010); Fridjonsson et al. (2011). Whilst all these studies lend important insight into MRI use for biofilm imaging and analysis, they cannot be mapped directly onto SuDS related research due to inappropriateness of scale and material.
Thus, despite a very recent surge in MRI application to flow-pollution processes prevalent in SuDS, the 20+ papers available do not consider the effect of biofouling on filter efficiency for metal removal at coarse-grain scales as required in SuDS design. Importantly, only 2 of the aforementioned studies consider a coarse-grained gravel media set-up appropriate to SuDS research:
Ramanan et al. (2012) successfully imaged superparamagnetic nanoparticle
___________________________________________________________________________________
103
transport through a course-grained quartz gravel media in a vertical column.
Also, Haynes et al. (2009) was able to image sedimentation of fine-grained sand within a gravel bed, an important aspect to gravel filter drains which encounter significant sedimentation of suspended solids from road runoff. Therefore, precedent exists for successful application of MRI to flow-sediment- pollution studies in coarse-grained media as necessary for gaining insight into SuDS filters, yet its application to biofouling of such a coarse-grain media is without such precedence. This therefore underpins the motivation of the present research.