Discovery of Photocatalysts for Hydrogen Production.pdf

Loading ...
Global Do...
News & Politics
Hydrogen, Fuel Cells, and Infrastructure Technologies
 FY 2003 Progress Report
Discovery of Photocatalysts for Hydrogen Production
D. Brent MacQueen (Primary Contact), Albert Hirschon, Theodore Mill, and Michael Coggiola
SRI International
333 Ravenswood Avenue
Menlo Park, CA 94025
Phone: (650) 859-5286; Fax: (650) 859-5286; E-mail: brent.macqueen@sri.com
Nobi Kambe and Timothy Jenks
NanoGram Corporation
46774 Lakeview Boulevard
Fremont, CA 94532
Phone: (510) 407-3232; E-mail: kambe@nanogram.com
DOE Technology Development Manager:  Roxanne Danz
Phone: (202) 586-7260; Fax: (202) 586-9811; E-mail: Roxanne.Danz@ee.doe.gov
Objectives
•
To design and construct tools for the high throughput screening of the properties of photocatalysts 
relevant to the photoelectrochemical generation of hydrogen.
•
To use the high throughput screening tools to identify materials with the appropriate energetics for 
water splitting.
•
To use the industrially relevant laser pyrolysis materials production capability of our partner, 
NanoGram Corp., to produce commercial scale quantities of photocatalysts.
•
Demonstrate that materials issues that are a hindrance to the commercialization of 
photoelectrochemical water splitting can be addressed through a materials discovery process.
Technical Barriers
This project addresses the following technical barriers from the Hydrogen Production section of the 
Hydrogen, Fuel Cells and Infrastructure Technologies Program Multi-Year R,D&D Plan:
• M. Materials Durability
•
N. Materials and Systems Engineering
•
O. Photoelectrochemical Efficiency
Approach
•
Develop tools for the high throughput analysis of material properties relevant to photoelectrochemical 
hydrogen generation.
• Work with existing nanoparticulate powder manufacturer to test materials with existing technology to 
produce kg/hr scale quantities.
•
Identify candidate materials for further study via high throughput screening.
Accomplishments
•
Constructed Solar Simulator.
•
Designed and constructed 25-cell photolysis analysis module.
1
Hydrogen, Fuel Cells, and Infrastructure Technologies
 FY 2003 Progress Report
•
Acquired diffuse reflectance instrument employing Ocean Optics Spectrometer.
•
Initiated evaluation of NanoGram NPMTM produced nanoparticulate titania.
Future Directions 
• Modify stable metal oxide semiconductors to shift bandgap to increase utilization of the solar spectrum 
and increase solar to hydrogen efficiencies.
•
Complete 25-cell module for electrochemical analysis of materials to provide band edge energy and 
photoaction spectrum analysis.
•
Incorporate hydrogen sensor technology developed for fuel cell safety applications into analysis 
modules for direct hydrogen determination.
•
Increase level of collaboration between other DOE funded laboratories via establishment of mutual 
database.

Introduction
The process whereby semiconductor materials 
absorb light and convert that light to electricity has 
been known for some time and is the cornerstone of 
the photovoltaic (PV) industry.  As an alternative to 
generating electricity, the electron/hole pair 
generated by the absorption of light by a 
semiconductor could be used to facilitate a chemical 
transformation, thereby storing the energy for later 
use.  This is the basis for the photoelectrochemical 
(PEC) generation of hydrogen from sunlight, water 
and a semiconductor-based photocatalyst.  In this 
process, electrons generated by the absorption of 
light are used to reduce water, generating hydrogen, 
and the holes are used to oxidize water, generating 
oxygen.  The hydrogen produced could then be used 
as a fuel in either electrolytic or combustion 
processes.  It is more efficient to generate hydrogen 
using PEC directly than to use a PV array to feed an 
electrolyzer.
The key materials requirements for efficient PEC 
hydrogen production are: A) a semiconductor with a 
bandgap relevant to the solar spectrum so as to 
maximize solar utilization, B) band edge energies 
relevant to the reduction and oxidation potentials of 
water to minimize energy mismatch, C) fast 
interfacial electron transport to minimize 
recombination and D) stability in the presence of an 
aqueous electrolyte.  The key technical barriers to the 
implementation of PEC hydrogen production relate 
to the fact that materials that satisfy all of these 
requirements, simultaneously, have not been found.  
This is not due to a conflict with these requirements 
and known laws of physics (there are none), but 
rather reflects the immature stage of this technology.  
In the earliest work on this technology, 
semiconductor materials developed for the PV 
industry were analyzed and found to be unsuitable 
due to their poor stability in aqueous electrolytes.  
Current work is under way with respect to the 
identification of materials specifically for the 
purpose of photoelectrochemical generation of 
hydrogen.
Approach
The use of high throughput techniques to speed 
the discovery process has been in place for a number 
of years; pharmaceutical companies were the first to 
invest heavily in the combinatorial synthesis and 
high throughput analysis concept.  The key to the 
concept is to test as many samples as possible as 
quickly as possible for a specific property rather than 
to do a complete characterization on a specific 
material or class of materials.  In this manner, 
candidates for further study can be culled from very 
large sample sets.  In our approach to develop tools 
for the high throughput screening of materials for 
properties relevant to PEC hydrogen production, we 
have designed and built a 25-cell module to analyze 
the photolysis products generated upon illumination 
of samples with a simulated solar spectrum.  The 
basic design incorporates a vacuum sealed sample 
module with a pressure transducer to monitor the 
pressure as a function of photolysis time, see Figure 
1.  The pressure transducer data is computer 
2
Hydrogen, Fuel Cells, and Infrastructure Technologies
 FY 2003 Progress Report
monitored, and at set pressure intervals, a Gillson 
sample handler (top right of Figure 1) that has been 
modified is triggered to sample the head space gas 
and feed it to a gas chromatograph for analysis.  The 
vacuum seal component of this design ensures that 
each cell is isolated, and the quick release and seal 
capabilities make loading quick and easy.
In the Hydrogen, Fuel Cells and Infrastructure 
Technologies Program (HFCIT) Multi-Year Program 
Plan, the need to have processes for the 
commercially relevant production of PEC materials 
is identified.  Part of the sample lots that we are 
investigating are produced by NanoGram Corp. 
using their nanoparticle manufacturing technology, 
NPMTM.  NanoGram has the process technology to 
produce these materials at rates of kilograms per hour 
and higher should a demand for these materials arise.  
Results
To date, the key results involve the design and 
construction of the 25-cell photolysis analysis 
module, which we have now fully engaged in the 
initial screening of sample sets of nanoparticulate 
titania provided by our partner, NanoGram Corp.  
The photolysis setup delivers a photon flux on the 
order of 80 mW/cm2 to the sample cells as 
determined by actinometry.  An interesting 
observation is that the pressure transducer response 
factor is slowly changing.  This was noticed during 
the course of a routine pressure testing in which we 
had initially observed very a very narrow distribution 
of pressure responses; however, after a number of 
experiments, the distribution of response factors 
appeared to increase, see Figure 2.  As such, we now 
routinely collect a calibration curve for the pressure 
transducers.  As shown in Figure 3, while the 
response factors of the various pressure transducers 
does vary considerably, the response of any given 
transducer is very linear over pressure ranges 
encountered during the normal course of operation.
A sample data set collected during the course of 
a photolysis experiment is shown in Figure 4.  
Contained with the 25 sample cells are 6 TiO2 
samples in triplicate, one Niobate in triplicate and 
four blanks containing a similar particulate that is 
known to be inactive.  The data set has been 
corrected for the photon flux for each specific cell 
and for the pressure transducer response.  From this 
Figure 1. Picture of 25-Cell Photolysis Analysis 
Module
Figure 2. Leak Test Results Indicating Range of 
Pressure Transducer Response Factor
Figure 3. Results of Calibration of Pressure Transducer 
Response Factor
3
Hydrogen, Fuel Cells, and Infrastructure Technologies
 FY 2003 Progress Report
experiment, the niobate sample resulted in the best 
efficiency and the highest pressure versus time plots, 
followed by an oxygen-deficient phase of titania 
which does extend the absorption spectrum into the 
visible.  The niobate, while still exhibiting 
efficiencies of less than 1%, is a material that we will 
be investigating further in various modified forms.  
To date, we have received over 50 different samples 
from NanoGram, with the majority of them being 
titania samples prepared over a significant range of 
particulate sizes, see Table 1.
In our initial screen of the TiO2 samples, a 
platinum catalyst was deposited onto the particle sets 
via the photoreduction of a platinum salt.  Samples 
have also been prepared using chemical reduction of 
impregnated platinum salts; however, these to date 
have not been screened.  In the titania series prepared 
with the photoreduced platinum catalyst, we see no 
statistical variation in the efficiency of hydrogen 
generation as a function of particle size, and in all 
cases efficiencies were in the range of 0.1% solar 
photons to hydrogen.  Interestingly, there does appear 
to be a trend developing with respect to the hydration 
of the titania.  These results are very premature, and 
considering the overall rather disappointing 
performance of these materials to date, it is doubtful 
that time will be spent analyzing them further with 
the platinum catalyst.  We are currently evaluating 
these materials with other catalysts.  We are also 
currently expanding the materials we are looking at 
to include some niobate structures.  These materials 
have shown evidence of water splitting in the past 
and the extension of their absorbance spectra into the 
visible relative to the titania samples, see Figure 5.
Conclusions
Our initial results, while far from complete or 
conclusive, indicate that the nanoparticulate titania 
samples provided by NanoGram Corp. exhibit very 
poor efficiencies for the conversion of photons of 
solar relevance to hydrogen.  In no case were 
conversion efficiencies greater than 0.3% observed in 
any of the titania samples.  This clearly indicates that 
we need to expand the composition of materials 
being explored.  With the relocation of one of 
NanoGram Corporation's laser pyrolysis instruments 
to the SRI campus, we will be able to facilitate this 
expansion of material candidates considerably.
Figure 4. Time Versus Pressure Results
Table 1. Partial TiO2 Sample List Indicating Particle 
Size and Surface Area
Figure 5.  Absorbance spectra generated from diffuse 
reflectance data for a titania, TOX 17, and a 
niobate semiconductor sample.
4
Hydrogen, Fuel Cells, and Infrastructure Technologies
 FY 2003 Progress Report
FY 2003 Publications/Presentations 
1. Brent MacQueen, "Photoelectrochemical 
Hydrogen Production", Invited Lecture, 
Hydrogen Workshop of the Global Climate 
Energy Project, March 2003, Stanford 
University, Palo Alto CA.
5