Elsevier

Energy

Volume 39, Issue 1, March 2012, Pages 63-73
Energy

Water droplet accumulation and motion in PEM (Proton Exchange Membrane) fuel cell mini-channels

https://doi.org/10.1016/j.energy.2011.10.023Get rights and content

Abstract

Effective water management is one of the key strategies for improving low temperature PEM (Proton Exchange Membrane) fuel cell performance and durability. Phenomena such as membrane dehydration, catalyst layer flooding, mass transport and fluid flow regimes can be affected by the interaction, distribution and movement of water in flow plate channels.

In this paper a literature review is completed in relation to PEM fuel cell water flooding. It is clear that droplet formation, movement and interaction with the GDL (Gas Diffusion Layer) have been studied extensively. However slug formation and droplet accumulation in the flow channels has not been analysed in detail. In this study, a CFD (Computational Fluid Dynamic) model and VOF (Volume of Fluid) method is used to simulate water droplet movement and slug formation in PEM fuel cell mini-channels. In addition, water slug visualisation is recorded in ex situ PEM fuel cell mini-channels. Observation and simulation results are discussed with relation to slug formation and the implications to PEM fuel cell performance.

Highlights

► Excess water in mini-channels from the collision and coalescence of droplets can directly form slugs in PEM fuel cells. ► Slugs can form at low flow rates so increasing the flow rate can reduce the size and frequency of slugs. ► One channel of a double serpentine mini-channel may become blocked due to the redistribution of airflow and pressure caused by slug formation. ► Correct GDL and mini-channel surface coatings are essential to reduce slug formation and stagnation. ► Having geometry changes (bends & steps) in the flow fields can disrupt slug movement and avoid channel blockages.

Introduction

Fuel cells produce electricity from a re-dox reaction of the fuel, hydrogen, with oxygen from the air. Fuel cells have the advantage of only producing water and heat as by products and are extremely efficient [1]. The PEM fuel cell is a low temperature electrochemical device that offers a promising, possibly green, alternative to traditional power sources and other fuel cell types in many applications without air polluting issues [2], [3], [4], [5]. At present some barriers still persist that are hampering PEM fuel cell system commercialisation. One of these barriers is water flooding, hindering the advance of PEM fuel cells, especially at low temperature operation [6].

At high current densities liquid water accumulates at the cathode. This is as a result of the ORR (Oxygen Reduction Reaction), remembering that the electrochemical result of combining hydrogen and oxygen at a temperature below 100 °C forms liquid water. Water can also be transported from the anode to the cathode, through the membrane, via electro-osmotic drag and local pressure, temperature and concentration gradients [7]. If a nafion membrane is used, it must be fully hydrated for it to be a good proton conductor. In order to maintain the membrane hydration level, especially at start up; the reactant and oxide gases are often fully humidified with water. When a PEM fuel cell accumulates too much water at high current density, about one-third of the electrode surface area may not be utilised [8], [9]. The phenomenon of flooding is a well established problem at the cathode electrode where both the catalyst layer and/or the GDL (Gas Diffusion Layer) may be mass transport-limited due to condensed water [10], [11]. This water flooding increases the internal resistance of the cell, blocks the ORR reaction, disrupts cell pressure and gas flow, making the cell voltage and current unpredictable and unrepeatable, reducing the PEM fuel cell performance dramatically. Many factors affect how water is transported from the catalyst layer to the flow plate and how droplet formation occurs, grows and ultimately how water flooding occurs within the PEM fuel cell flow channels. These factors include the membrane, flow plate or flow field channel design, airflow rate, temperature, power density, gas humidification as well as PEM fuel cell orientation.

If the cathode layer or GDL floods, water must be transported from the electrode, through the GDL and into the flow plate flow channels and then exhausted out of the cell to mitigate the flooding problems. The water transport mechanism from the GDL into the flow plate channels has been explained by two theories; converging capillary tree water transport mechanism proposed by Nam and Kaviany [12] and channelling liquid water transport mechanism proposed by Litster et al. [13]. There are discrepancies between both theories; however, both theories concur that water droplets can form on the GDL and that these droplets, mainly originate from the electro-catalyst layer. These droplets join and grow in the GDL and squeeze out into the channels due to capillary action within the GDL, its hydrophobic nature and due to pressure and temperature forces in the flow plate channels [14], [15], [16], [17].

Many researchers have observed and modelled water droplet formation and emergence from GDL pores [18], [19]. Kimball et al. [22] measured the critical hydrostatic pressure head for liquid water breakthrough for various GDL materials. Kumbur et al. [20] employed an ex situ flow channel apparatus in order to study droplet formation and instability. Park et al. [21] modelled fluid flow through GDLs where it was concluded that a thin GDL with small porosity results in good electrical conductivity; however efficient mass transport requires large pores. Many authors including, Yang et al. [23], Hakenjos et al. [24] and Gao et al. [25] describe preferential location emergence of droplets from the GDL relating it to temperature distribution or capillary action within the GDL. Zhu et al. [26] modelled liquid water entering a PEM fuel cell channel through a GDL pore. A 2-D VOF model was employed to view the effects of flow channel size, droplet coalescence and pore size on the emerging water droplet dynamics. It was found that in large micro-channels (0.5 mm) droplet deformation slows down and droplet breakup may not occur. This can result in a film of water forming against the GDL downstream of the pore. Researchers have also produced water flooding mitigation methods and models. Tüber et al. [27] visualised liquid water transport in the cathode gas channel of a transparent PEM fuel cell at low operating temperatures (30 °C). They found that using a hydrophilic cathode GDL resulted in increased current density, which they attributed to a more uniformly hydrated membrane. This was also noted by Ge and Wang [28] who visualised water droplet formation in the anode flow channels. They observed that droplets tended to form on the gas channel walls when a hydrophobic GDL was employed, whereas a hydrophilic GDL tended to wick water from the channel into the GDL. Quan et al. [29] also studied the water management in a PEM fuel cell flow channel. This study focused on the effects of channel hydrophilicity, channel geometry, air inlet velocity and pressure drop in relation to water behaviour. It was concluded that sharp corners inside the channel with the aid of hydrophilic surfaces may aid in water transport and due to increased spreading of the liquid, pressure drop is increased in the flow channel. Liu et al. [30] investigated the liquid water accumulation in PEM fuel cells cathode flow channels with three different flow field configurations: parallel, interdigitated, and cascaded. At low operating temperatures (25 °C) and ambient pressure, they observed that the parallel flow field was the most unsuitable flow field design for water removal, resulting in the worst performance of the PEM fuel cell. In a study by Zhu et al. [31], the effect of micro-channel geometry on water droplet dynamics in a PEM fuel cell using a 3-D VOF model, was investigated. They compared many different micro-channel designs; rectangular, trapezoid, upside down trapezoid, triangular, rectangular with curved bottom wall and semicircular with respect to evolution and motion of the droplets, flow resistance, saturation and coverage ratio. They concluded that the geometry and the micro-channels wet-ability drastically affect water droplet movement which should help design better flow channels for more effective water removal. Carton and Olabi [32] performed a Design of Experiment (DOE) study on three PEM fuel cell flow plate configurations. It was concluded that over all the serpentine flow plate performed best, due to its continuous flow design water flooding was mitigated. Li et al. [33] also concluded that the serpentine plates offer the best results for water removal, since it ensures the removal of water produced from a cell with acceptable parasitic load. Ous and Arcoumanis [34] noted that increased airflow rate prevented droplet formation but reduced current due to membrane dehydration. Again, Weng et al. [35] confirmed the beneficial effects of high cathode gas flow rates for water removal; however un-humidified cathode gas streams at high stoichiometry resulted in membrane dehydration.

Overall, however, only a few researchers have investigated slug movement in flow channels and this area has not been analysed to the same degree as droplet formation and emergence from GDL pores. Hussaini and Wang [36] presented images corresponding to the most common flow patterns in operational PEM fuel cells. These include; single phase flow, droplet flow, film flow and slug flow as shown in Fig. 1. The authors explain that these flows can evolve from each other resulting in slug flow, which was a consequence of film flow growth. Lu et al. [37] also observed three different types of water to airflow regimes in their study, slug flow, annular/film flow and mist flow. They noted that slug flow and intense annular flow cause an increased pressure drop due to liquid water build-up, which is a key cause of flow misdistribution that dramatically reduces the PEM fuel cell performance and durability.

Zhou et al. [38] studied water behaviour in the cathode side of a PEM fuel cell, with a serpentine channel using VOF. Water droplets and films were introduced into the channel at varying positions, to simulate different operational conditions of the PEM fuel cell and high airflow rates (10 m s−1) were used. Detailed results show droplet evolution, breakup and movement of water films through the serpentine micro-channel. However, Ous and Arcoumanis [34] viewed the formation of water droplets emerging from an operational PEM fuel cell (Fig. 2). A transparent proton exchange membrane fuel cell was used to visualise the water droplet formation during its operation. Visualisation results show that water accumulates first in the middle flow channels and that no accumulation takes place at the bend areas. Droplets were observed to appear and then shrink on the GDL, which they attributed to their increased cross-sectional area in the direction of the airflow, which may push the droplet back into the GDL. The droplets grew larger and then adhered to the channel walls. It is shown that after 95 min of operation droplets can join to fill the channel, as shown in Fig. 2. Measurement of the fuel cell current during water production showed that the current gradually declined as more water filled the channel.

Water flooding analysis, in literature, has mainly focused on droplet and micro-droplet release from the GDL and their interaction within the flow field channels. However, to design effective flow channels and to aid the mitigation of water flooding, water droplets after their emergence from the GDL and consequential interaction with each other and the flow channels must be investigated and analysed. It has also been shown from literature that water, from the GDL, has preferential location emergence and does not emerge in a film form, but discrete droplets on the GDL, if a hydrophobic GDL is used. These droplets can then join to form a slug. The definition of a slug in this study is a large water droplet that adheres to either the GDL or channel wall and moves in airflow. In this study, a two-phase flow model is developed, viewing three different scenarios, where the coalescence of droplets and movement of water slugs in flow field mini-channels are investigated. To simplify the model water droplets are introduced into the channel inlet. This model, however, can represent any region of the flow plate where slugs may occur. In addition with the aid of imaging techniques, visualisation of water droplets and slugs are recorded in ex situ flow channels to ensure valid results from the model. The implication of slugging on flow plate design and PEM fuel cell performance is discussed.

Section snippets

Volume of fluid model

In the present study suitable CFD (Computational Fluid Dynamic) software using VOF (Volume of Fluid) was applied in order to simulate fluid motion in mini-channels without heat transfer. VOF is a surface-tracking technique applied to a fixed Eulerian mesh. It is designed for two or more immiscible fluids, where the position of the interface between the fluids is of interest. In the VOF model, a single set of momentum equations are shared by the fluids, and the volume fraction of each of the

Experimental setup

A detailed review of liquid water visualisation in PEM fuel cells was recently conducted by Bazylak [43]. Many liquid water visualisation techniques including NMR (nuclear magnetic resonance), beam interrogation and direct optical imaging techniques were mentioned. The reviewer concluded that even with the difficulties posed by direct optical imaging such as opaque materials or in/ex situ apparatus and limitations of magnification and speed of the camera used, direct optical visualisation (used

Model validation

A time sequenced VOF model was simulated to view a slug movement around the double serpentine channel bends, as shown in Fig. 6(a). To ensure that the model matched reality, an experiment was performed to view slug movement around the bend of a double serpentine channel Fig. 6(b). The model and experiment shows three time instances of the slug movement as shown in Fig. 6. As the slug is pushed with the air it begins to deform Fig. 6(1). Depending on the surface treatment of the channel and or

Discussion

If PEM fuel cells are operated below 100 °C liquid water is produced in the cell. It is noted that if nafion based membranes continue to be used, good PEM fuel cell performance depends on good water management [44]. Water flooding in PEM fuel cells is a major issue and mitigation methods and/or techniques will ensure the full potential of these electrochemical devices.

The CFD VOF model was used in this study in order to investigate the interaction of the GDL and mini-channels with water

Conclusions

In this study, using CFD modelling techniques a two-phase flow model was used to investigate the coalescence of droplets and movement of slugs in flow field mini-channels and with the aid of imaging techniques, visualisation of the water droplets and slugs were recorded in an ex situ apparatus. The implication of slugging on flow plate design and PEM fuel cell performance were discussed.

It is concluded that:

  • Excess water in mini-channels from the collision and coalescence of droplets can

Acknowledgement

The authors would like to thank the School of Engineering and Environmental Sciences in Wels Austria for use of their imaging devices.

References (47)

  • A. Hakenjos et al.

    A PEM fuel cell for combined measurement of current and temperature distribution, and flow field flooding

    Journal of Power Sources

    (2004)
  • B. Gao et al.

    Visualization of unstable water flow in a fuel cell gas diffusion layer

    Journal of Power Sources

    (2009)
  • X. Zhu et al.

    Dynamic behaviour of liquid water emerging from a GDL pore into a PEMFC gas flow channel

    Journal of Power Sources

    (2007)
  • K. Tüber et al.

    Visualization of water buildup in the cathode of a transparent PEM fuel cell

    Journal of Power Sources

    (24 Nov 2003)
  • P. Quan et al.

    Numerical study of water management in the air flow channel of a PEM fuel cell cathode

    Journal of Power Sources

    (10 Jan 2007)
  • Z. Liu et al.

    A two dimensional partial flooding model for PEMFC

    Journal of Power Sources

    (25 Aug 2006)
  • X. Zhu et al.

    Numerical investigation of water droplet dynamics in a low-temperature fuel cell microchannel: effect of channel geometry

    Journal of Power Sources

    (01 Feb 2010)
  • J.G. Carton et al.

    Design of experiment study of the parameters that affect performance of three flow plate configurations of a proton exchange membrane fuel cell

    Energy

    (2010)
  • H. Li et al.

    A review of water flooding issues in the proton exchange membrane fuel cell

    Journal of Power Sources

    (15 Mar 2008)
  • T. Ous et al.

    Visualisation of water droplets during the operation of PEM fuel cells

    Journal of Power Sources

    (08 Nov 2007)
  • F. Weng et al.

    Study of water-flooding behaviour in cathode channel of a transparent proton-exchange membrane fuel cell

    Journal of Power Sources

    (03 Jul 2006)
  • I.S. Hussaini et al.

    Visualization and quantification of cathode channel flooding in PEM fuel cells

    Journal of Power Sources

    (2009)
  • Z. Lu et al.

    Water management studies in PEM fuel cells, Part II: ex situ investigation of flow maldistribution, pressure drop and two-phase flow pattern in gas channels

    International Journal of Hydrogen Energy

    (2009 May)
  • Cited by (232)

    View all citing articles on Scopus
    View full text