Skip to main content

Hydrogen Contaminant Risk Assessment

Final Report Summary - HYCORA (Hydrogen Contaminant Risk Assessment)

Executive Summary:
The main objective of HyCoRA project has been to provide information to reduce the cost of hydrogen fuel’s quality assurance (QA). It has also provided recommendations for revision of existing ISO 14687-2:2012 standard for hydrogen fuel in automotive applications.

For developing the strategy for cost reduction for hydrogen fuel QA, a hydrogen quality risk assessment has been used to define the needs for hydrogen impurity gas analysis, system level PEMFC contaminant research and purification in hydrogen production, especially by steam methane reforming (SMR) with pressure swing adsorption (PSA).

Hydrogen fuel QA affects directly to the market penetration of the fuel cell electric vehicles (FCEVs). Unsuccessful QA would lead to vehicle incidents and hinder the public acceptance of the technology. On the other hand, unnecessarily tight fuel QA induces costs that are directly translated into the fuel cost for the customer. Without reasonable pricing, even good technology will not be adopted by the users.

During HyCoRA project, it has been shown that the limits of formaldehyde (HCHO) and formic acid (HCOOH) in the ISO 14687 standard could be relaxed. The FC measurement test system utilized for H2 fuel impurity measurements is state of the art. The automotive-alike approach to H2 recirculation, implemented into a test stations by VTT, has been omitted also by other groups (e.g. LANL). During HyCoRA project, three sampling campaigns from the hydrogen refuelling station’s (HRS) nozzle have been completed, fuel composition and impurity concentrations analysed. The technical data gathered from the three H2 fuel-sampling campaigns is unique and public and has induced much interest. The HRS sampling strategy has been widely accepted and successful sampling campaigns culminated in Annex I to 19880-1. In addition, a probabilistic risk assessment model for determining QA needs has been built and released as open Matlab code for further exploitation.

HyCoRA project results have been disseminated in various international conferences and workshops, and presented in TC 197 meetings for WG 24, 27 and JWG 7.
Project Context and Objectives:
The FCEVs possess a great promise for decarbonisation of the traffic. However, their penetration to the market requires a quality-controlled and affordable hydrogen fuel. Thereby proper hydrogen fuel QA is required, to avoid harmful substances to enter the fuel cell. All the incidents affect the public acceptance, no matter whether they are large incidents, where you need to stop driving and get your vehicle towed to the maintenance, or small incidents, where your vehicle’s maximum power output is limited until the next recovery. On the other hand, many of the analysis, in- and offline, can be costly especially for the low levels of impurities considered in hydrogen fuel. Both, public acceptance and the cost of H2 fuel have an impact on the market penetration of FCEVs.

The overall objective of HyCoRA has been to reduce the cost of H2 fuel’s quality assurance and to provide recommendations for revision of existing standard for H2 fuel in automotive applications, ISO 14687-2:2012.

The sub objectives:
- Complete the current knowledge by identifying the impurity limits of PEMFCs for various poisonous species under actual automotive drive cycles
- Compare the relevance of single cell impurity testing, when compared to results obtained using miniature automotive systems
- To identify and develop novel methods for the analysis of trace-level impurities in hydrogen in order to provide robust quality assurance measurements of hydrogen fuel for use in polymer electrolyte membrane fuel cell (PEMFC) applications
- To simplify and reduce cost of analysis by reducing the number of analytical techniques required, partly through the establishment and validation of a pre-concentration device
- To validate the performance and accuracy of methods for the quality assurance of hydrogen fuel which are currently used by industry
- To assess the hydrogen fuel quality of HRS's in Europe
- Sampling of hydrogen from HRS nozzle
- Analysis of samples collected from HRS in accordance with prevailing standards
- Survey on quality of hydrogen from producers as input to risk assessment
- To construct a probabilistic risk assessment model that integrates the data on hydrogen quality variation and correlations between impurity concentrations, hydrogen impurity analysis methods and instrumentation, and the susceptibility of fuel cells to hydrogen fuel contaminants produced in WP3, WP2 and WP1.

The objectives of HyCoRA were met.
Project Results:
3.1 Determination of susceptibility of hydrogen contaminants for automotive applications

In HyCoRA, a more complete overview of the real susceptibility of various contaminants in automotive operation has been achieved. Measurements with automotive alike FC system were needed for determination of correct limits of HCHO and HCOOH. The results showed, that 2.0 ppm HCHO and 20 ppm HCOOH, that are 200 times and 100 times the current limits in ISO 14687-2:2012, have a very small effect. The effect is notably smaller than the effect of CO. An approximately 10 mV voltage drop was seen in four hours of operation with 2.0 ppm HCHO in pure 6.0 grade H2, and 6 mV voltage drop in four hours with 20 ppm HCOOH. In reference measurements with 2 ppm CO, the cut-off limit of 50 mV was reached in 55-72 min.

The current limit in ISO 14687 for HCHO is 0.01 ppm, for HCOOH 0.2 ppm and for CO 0.2 ppm. According to the results of HyCoRA, the limits for HCHO and HCOOH could be safely relaxed at least with one decade.

It has been concluded, that hydrogen impurity measurements should be conducted with FC stacks instead of single cells. VTT’s in-house build miniature automotive test bench for 1 to 2 kW stacks, Figure 2. There are plenty of impurity measurements conducted with single cells, which may not be representative for automotive use. Even if the automotive-alike single cells in relevant operating conditions could be utilized, the hydrogen recirculation becomes more demanding to realize with single cells, disturbs water balance and leads to notably larger gas volume at the anode side when compared with real automotive system configurations. In addition, online gas sampling is impossible to perform without altering notably the gas composition in the recirculation loop.

However, with new impurities and aggressive irreversible impurities it might make sense to perform (first phase) studies with single cell, due e.g. to the cost issues.

In HyCoRA, it has been demonstrated that the impact of CO is mitigated by the anode gas recirculation, the voltage cycling and SU/SD cycling. The impact of 1 or 2 ppm CO is significant for low loaded MEA, but not with high loaded MEA.

Impact of H2S in single cell and with fuel cell dynamic load cycle (FC DLC) has been studied. Even if voltage cycling had been shown to result a voltage plateau for the impact of H2S, there was no positive effect on H2S tolerance due to the FC DLC seen in HyCoRA experiments. Nevertheless, OCV showed some cleaning effect of the anode, but OCV increases also the degradation of the MEA.

Combined contamination with H2S and CO showed that while H2S has a clear effect on the performance for pure H2, it prevents the full coverage of CO and therefore allows higher voltage under H2 + 1 ppm CO.

3.2 Development and validation of novel analytical methods for hydrogen quality

One of the main limitations for methods for hydrogen quality is analytical sensitivity. Pre-concentration of samples is one way to reduce the number of analytical techniques required. An alternative approach to pre-concentration is cryofocusing. In HyCoRA, palladium membrane hydrogen separation has been tested. In addition, VTT has looked into pre-concentration of formaldehyde and subsequent analysis with GC-FID, Figure 3. SINTEF has incorporated cryofocusing into a GC/MS setup. They have analysed samples collected from HyCoRA sampling campaign and the results indicate good sensitivity, especially with respect to hydrocarbons. Pre-concentration and cryofocusing efforts reported in HyCoRA deliverable D2.4.

Protea LTD has designed and tested the FTIR and MS instrument, specifically designed for the combination of FTIR and QMS in a single instrumental system that performs the measurement of all trace impurities in H2, Figure 4. The detection limits on the analyser built during the project meet or exceed the ISO/DIS 14687-2 requirements for nearly all gases. The detection limits of H2O and total sulphur are not low enough to meet the ISO requirements. The addition of a QCL laser measurement system within the same analyser would improve the H2S and H2O detection, but the total sulphur measurement meeting the 0.004 ppm level requires a separate configuration with a path longer than 20 meters.

The design of the FTIR and MS analyser, added a QCL analyser for H2S, show that the list of impurity constituents can be covered analytically by three analytical techniques. This is a significant improvement compared with the six analytical techniques currently used by Smart Chemistry, subcontracted to analyse the 28 HRS samples collected in HyCoRA.

3.3 Assessment of hydrogen quality variation in hydrogen refuelling stations

Three sampling campaigns of commercial hydrogen refuelling stations (HRSs) have been performed in HyCoRA. The first sampling campaign, in December 2014, concentrated on the feedstock. Eight samples with feedstock of chlor-alkaline, water electrolysis, SMR as well as compressed and liquefied H2 were collected from Germany and Norway. The second sampling campaign was conducted in June 2016, and the HRSs were selected with the aim of targeting as newly commissioned stations as possible. The third sampling campaign was conducted in spring 2017. No specific objective for selection of HRSs was obvious based on previous results, and HRSs in Scandinavia were chosen. A total of 28 gas samples and 14 particulate samples were collected and 42 samples analysed in accordance with standards. The unique data and full reports of the sampling campaigns are public, HyCoRA deliverables D3.2 and D3.3.

All the gas samples have been collected at the nozzle, at representative pressure and velocity. The sampling was realized with Linde’s Qualitizer equipment and with Linde’s Spectraseal lined cylinders.

Hydrogen fuel quality was noted generally good, only few violations observed since 2012. Significant impurity levels observed for H2, He, O2, CO2, H2O, THC, C4Cl4F6. Impurities did not correlate with H2 feedstock, neither was correlation between commissioning date and fuel quality found.
In addition, two different ways to collect simultaneously particulate and gas samples were tried out. The samples were collected with a HYDAC PSA70 Sampling equipment that was placed either before or after the Linde sampler. The results indicate that sampling of gas and particulates should be performed separately.

3.4 Risk assessment of hydrogen quality assurance failure

The HyCoRA risk and cost model is a probabilistic risk assessment model for hydrogen fuel quality, implemented in Matlab code. Full documentation with the Matlab code is openly available, HyCoRA deliverable D4.3. The model allows assessing the effect of fuel quality control (QC) measures introduced in the fuel delivery chain on the risk of degraded FCEV performance caused by contaminants in the fuel, Figure 6. Furthermore, the model allows calculating the overall cost associated with the measures, comprising the investment and operating costs of the QC measure and the damage costs from vehicle incidents expected still to occur, Figure 7. The model applies Monte Carlo simulation to deal with and process the various sources of uncertainty involved in such assessment.

The model has been limited to PEMFCs in automotive application, and automotive grade hydrogen fuel. During the project, the work has been further limited to centralised production from natural gas (NG) using the SMR with PSA process as the production-purification pathway. The implementation has been further limited to CO as the single contaminant accounted, even though the effects of CO2, Cl and S are included in the modelling as other relevant gaseous impurities.
Potential Impact:
HyCoRA project has provided information to reduce the cost of hydrogen fuel quality assurance and recommendations for revision of existing standard for hydrogen fuel in automotive applications, ISO 14687-2:2012.

In HyCoRA, a more complete overview of the real susceptibility of various contaminants in automotive operation has been achieved. The measurements with 2 ppm HCHO and 20 ppm HCOOH show, that their limits in ISO 14687 standard could and should be relaxed. Presentation of the HCHO and HCOOH project results have been essential for the introduction of a total budget for these and CO into the standard, but it is not yet in the draft of CD 14687.

With the influence from the FC measurement results conducted in HyCoRA, recirculation is becoming state-of-the-art and other groups worldwide are adapting to the system level studies with H2 recirculation for the FC impurity measurements.

HyCoRA project has resulted a very unique data set from three sampling campaigns of hydrogen fuel from the HRS nozzle. Before the results of HyCoRA, there was no public data available on HRS hydrogen fuel quality. In addition to gas sampling, particulate sampling was conducted during two latter campaigns. It has been shown in HyCoRA, that most of the HRS meet the H2 fuel quality standard.

The HRS sampling strategy of HyCoRA has been widely accepted and the sampling strategy has been written into Annex I to ISO standard 19880-1.

HyCoRA project and project’s results have established much of knowledge on hydrogen fuel quality and impurity impacts and the results have been disseminated in various international conferences and workshops, and presented in TC 197 meetings for WG 24, 27 and JWG 7. HyCoRA consortium has organised three OEM workshops on the topic “Hydrogen fuel quality assurance for PEM fuel cells”. Several scientific publications are in preparation. To facilitate further the exploitation of the project results, almost all of the deliverables are public and can be found from the project’s website.
List of Websites:
Project website:

Jaana Viitakangas, coordinator
Jari Ihonen
VTT Technical Research Centre of Finland LTD