nano 2007-01-21 06:26
Nanostructured Membranes: A New Class of Protonic Conductor..
[align=center][b]Nanostructured Membranes: A New Class of Protonic Conductor for Miniature Fuel Cells
SC5E*_Q d ?6H#F
dr5G KV'n!r
Bernard Gauthier-Manuel and Tristan Pichonat[/b][/align][align=left]
g I)Gy"E `
{n:lin a3g
********************************************************
h @^x^Zbl
[url=http://www.azonano.com/details.asp?ArticleID=1303#_Abstract][color=#0000ff]Abstract[/color][/url]c/g5E$D2nW{
[url=http://www.azonano.com/details.asp?ArticleID=1303#_Background][color=#0000ff]Background[/color][/url] Yb6V,Ud~#y qv7m
[url=http://www.azonano.com/details.asp?ArticleID=1303#_Methods_and_Materials][color=#0000ff]Methods and Materials[/color][/url]
'PX@+bN*bT)j U
[url=http://www.azonano.com/details.asp?ArticleID=1303#_Results_and_Discussion][color=#0000ff]Results and Discussion[/color][/url]0q&T9A~r;q)G:v6y@
[url=http://www.azonano.com/details.asp?ArticleID=1303#_Conclusions][color=#0000ff]Conclusions[/color][/url]
_1l3{\7c8h7w!cv
[url=http://www.azonano.com/details.asp?ArticleID=1303#_References][color=#0000ff]References[/color][/url]
z2t9rnx(C j9z8y
[url=http://www.azonano.com/details.asp?ArticleID=1303#_Contact_Details][color=#0000ff]Contact Details[/color][/url]
(zp"q |+{
********************************************************
U2N/d `1b
y
AeOj"x$N
[b]Abstract[/b]eD
B%P&?*y
$@!e Hv2o)I
In this paper a new way to realize miniature fuel cells (FC) is demonstrated using nanostructured porous silicon (PS) as a protonic conductor. This solution involves the chemical grafting of molecules containing ionizable groups on the pore walls to mimic the structure of an ionomer, such as Nafion®, usually used to ensure the proton conductivity of proton-exchange membrane (PEM) fuel cells. Using this technique an inorganic, structurally stable, proton conducting membrane is produced with many optimizable parameters such as the pore size and the pore structure of the membrane and the nature of the grafted molecules. This potentially disruptive technology allows the production of low cost small fuel cells able to produce significant electrical current without the disadvantages related to the use of ionomers. The performances obtained by these materials is of a better quality, compared to similar fuel cells using Nafion®.u.A$TL3Lel
F/MX6n dX&xO
[b]Background[/b];a'XS e@ `F9E!i;]
X&wL4X#vj%y
Nowadays the design of energy sources able to power portable devices like cellular phones, laptop computers or nomadic sensors networks is a challenge. These devices are presently powered with batteries that limit their autonomy and require human intervention and electrical energy sources to recharge. Moreover they generate wastes incompatible with their proliferation. In the next 10 years the international technology roadmap for semiconductors [1] forecasts a decrease of the voltage required to power the working of microcircuits toward 0.6 V. The use of miniature fuel cells (FC) appears an attractive way to power portable electronics with a clean and refillable energy source. This is one of the reasons to explain the intense activity presently occurring in the FC research field. Among all kinds of FC only two are really suitable for miniaturization. The limitation principally comes from the working temperature required to be lower than 100°C. One of them is the proton exchange membrane (PEM) FC. A key element of a PEM FC is the membrane that must have high conductivity for protons and be impermeable to all other present species (H2, O2, water, any other fuel, etc).
2m'P|f#{3Gg
_6M
O E-U6hln.t~
State-of-the-art miniature FC [2-6] generally use ionomer films to conduct protons from the anode, where hydrogen is consumed, to the cathode producing, with the reduction of oxygen, water, electrical current and heat. At the present time, the best conductivity (0.08 S.cm-1) is reached by Nafion® perfluorosulfonated membranes. However the high cost and the geometric instability during hydration are only some of the severe constraints of such polymers. j3` j4Q b
D aE`!V
The nature of the proton conductive process in an ionomeric membrane such as Nafion® is not still completely understood. The present consensus [7] is to describe a Nafion® membrane as a skeleton of hydrophobic chains including connected hydrophilic domains containing water molecules. The diameter of the connecting channels is about 3 nm. The low stiffness of this skeleton is responsible for the swelling of the membrane with hydration in response to the molecular interactions. sj G[^ z4r
*s@g*[w
TM$mk
Here, a potentially disruptive technology is demonstrated that allows a new way to realize the protonic conduction function in a small size FC.
mk mw8j8R5t3@P
1Dv x-|7i czX.G$l:i
Methods and MaterialsThe idea developed in this work is to try to reproduce the molecular structure of Nafion®films using harder inorganic materials. To be fully compatible with microelectronics processes and standard microfabrication techniques the FC will be integrated on silicon substrates. The membrane is then made of [color=black]porous silicon (PS) [/color]and direct grafting of proton conducting molecules on the surface of the pores ensures the required conductivity.
l8Td7C9V g o2vA
z1k9_#[
i&X!{w
The first stage is the realization of the silicon membranes using classical photolithography and wet chemical etching in KOH of (100) oriented silicon wafers. Masking is obtained with a sputtered Cr-Au layer (Cr layer of 15 nm thick and Au layer of 800 nm thick). A previous thermal oxidation of the silicon wafer ensures the electrical insulation of the membrane[color=black]. The[/color] memb[color=black]ranes thickn[/color]ess is fixed to 50 µm by adjusting processing time and temperature. Collective processing allows us to obtain simultaneously 69 membranes on a 4” wafer.
WDd[P,W_f
U.I6y0U*m N
Silicon membranes are then made porous, in a second stage, by anodization in a double-tank cell [8] conceived by AMMT GmbH and consisting of two half-cells in which Pt electrodes are immersed. The silicon wafer separates and isolates the two half-cells. The electrolyte used for anodization is an ethanoic-HF solution (50 % of pure ethanol and 50 % of a 48 % HF solution). Anodization is carried out in the dark at constant current. With phosphorus-doped 0.012-0.014 ohm.cm n+-type silicon wafers and a current density from 18 to 36 mA.cm-2, we obtain pores from 6 to 10 nm diameter and a porosity of about 50 % [9]. Across section of the porous silicon membrane obtained is sketched on the figure 1[color=blue].[/color]
#B s r"R&{:vE2A
[/align][color=#0000ff][/color][table=98%][tr][td][align=center][img=400,174]http://www.azonano.com/work/oS1JMS2e7S93s7CFP7f4_files/image001.jpg[/img] [/align][/td][/tr][tr][td][align=center][b]Figure 1.[/b] Schematic cross-section view of the porous silicon membrane. Silicon is in grey, porous silicon is in pink.[/align][/td][/tr][/table]The previous deposition of the Au layer allows the localization of the porosity on silicon membranes during anodization. Indeed a simple LPCVD Si3N4layer generally used as the masking layer for localized PS (see for example [10]) would not allow[color=black] lengthy an[/color]odization as precious metals (Au, Pt and A[color=black]g) allow. T[/color]hesilicon oxide layer under the Cr-Au layer avoids any parasitic formation of PS with the possible internal current generated between silicon substrate and the metallic layer [11]. Once anodization is achieved, membranes are rinsed into an oxidizing bath (Le Vert decontamination solution from Prevor) neutralizing the HF solution. Then several deionized water baths and isopropyl alcohol are used for rinsing and reducing stress into pores. Membranes are finally dried in ambient air. PS membranes with a silica surface layer are obtained. The characterization of these porous membranes is made by FESEM (Field Effect Scanning Electron Microscope) imaging and a typical view of the cross section of the membrane is represented in figure 2.[table=98%][tr][td][align=center][img=400,279]http://www.azonano.com/work/oS1JMS2e7S93s7CFP7f4_files/image002.jpg[/img] [/align][/td][/tr][tr][td][align=center][b]Figure 2. [/b]FESEM cross section view of a n+-type porous silicon membrane. Channels have an average diameter of 10 nm.[/align][/td][/tr][/table]As proved by conductivity measurements [12] and FESEM imaging, only a few channels are totally opened with this anodization process. Indeed due to the inhomogeneity of the wafer thickness, when the first channels open on the back side of the membranes, the current goes through these opened pores and the anodization no longer continues on the other pores. To solve this problem, a short reactive ion etching (R.I.E.) process is employed using SF6and O2gases for silicon etching on the back side of the membranes in order to make sure that all the pores are opened [12]. This process etches about 2 µm thick in 3 min. which is necessary to open the whole back side porosity. The characterization of opened pores is carried out by conductivity measurements in a 3 % hydrochloric acid electrolyte solution.
W-Y f4P(@
0z3`a {0Kr-`;r
To ensure the protonic conductivity and contrary to a previous solution consisting of filling pores with a Nafion® solution [12], a grafting method is describe here where the internal surface of the pores are covered with silane molecules bearing acid functions. This process is designed to mimic the supposed structure of Nafion® membranes and obtain significant proton conductivity. The N-[(3-trimethoxysilyl)propyl]ethylenediamine triacetic acid in the form of trisodium salt, commercially available from United Chemical Technologies Inc (UCT), was chosen[color=black] for the[/color] first investigations. As the surface of PS is covered by an oxide layer, the classical process of silanization can directly be used. The first step consists of creating silanol functions (Si-OH) at the surface of PS. A soft process involving UV ozone cleaner has been successfully implemented. This process creates the desired functions without geometrical modifications contrary to a previous wet process needing an immersion into “Piranha” solution (mixture of a 80 % solution of pure sulfuric acid with 20 % of a 33 % aqueous solution of hydrogen peroxide) which induced membrane deformations. The grafting of silane molecules is then realized by immersing the hydrophilic porous membranes into a 1 % solution of the acid silane in ethanol for 1 hour (time empirically determined with conductivity measurements) at room temperature and ambient air. Figure 3 shows a molecular simulation of the process.[table=98%][tr][td][align=center][img=389,420]http://www.azonano.com/work/oS1JMS2e7S93s7CFP7f4_files/image003.jpg[/img] [/align][/td][/tr][tr][td][align=center][b]Figure 3. [/b][color=black]View of[/color] the molecular scale of silane molecules compared to a 6nm diameter PS pore. One of the molecule is grafted on the surface. Si (yellow), O (red), H (white), C (grey),and N (blue).[/align][/td][/tr][/table]In order to replace -Na endings by -H endings to get the real carboxylic behaviour for the grafted function, membranes are immersed for 12 hours in a 20 % solution of sulfuric acid, then fully rinsed in deionized water. To complete the FC assembly, electrode and catalyst layers are added to the membrane. E-tek electrodes composed of a carbon conducting cloth filled with platinum (20 % Pt on Vulcan XC-72) were used as a H2 / O2 catalyst. A 1 µl drop of a 5 % Nafion®-117 solution provides, on each side, after the evaporation of the solvant, a proton-conducting link between the electrodes and the membrane. This amount of Nafion® solution used as glue is too small to fill the pores and does not contribute to the proton conductivity in the membrane. As the membrane borders are covered with a Cr-Au layer to collect current, it is important to make sure that the electrodes and the membrane plated borders are in contact.1?k*c~;KzY@
#N
y)d4A
jry&FY5C
:downloads :nst1 :pdf
suifeng0202 2007-05-12 01:01
感谢楼主提供的好文章！