HAZARDS ASSOCIATED WITH THE USE OF ACTIVATED CARBON CRYOGENIC GAS PURIFIERS

HAZARDS ASSOCIATED WITH THE USE OF ACTIVATED CARBON CRYOGENIC GAS PURIFIERS IGC Doc 43/14 Revision of Doc 43/07 EUROPEAN INDUSTRIAL GASES ASSOCIATION AISBL AVENUE DES ARTS 3-5 B 1210 BRUSSELS Tel : +32 2 217 70 98 Fax : +32 2 219 85 14 E-mail : info@eiga.eu Internet : www.eiga.eu

Doc 43/14 HAZARDS ASSOCIATED WITH THE USE OF ACTIVATED CARBON CRYOGENIC GAS PURIFIERS PREPARED BY : Witold Balczarczyk Andrew Clark Bernard Courtoy Ken Kovak Andrea Mariotti Dirk Reuter Inaki Uriarte Andy Webb The Linde Group Air Products Air Liquide Air Products SOL Messer Group Praxair España EIGA Disclaimer All technical publications of EIGA or under EIGA's name, including Codes of practice, Safety procedures and any other technical information contained in such publications were obtained from sources believed to be reliable and are based on technical information and experience currently available from members of EIGA and others at the date of their issuance. While EIGA recommends reference to or use of its publications by its members, such reference to or use of EIGA's publications by its members or third parties are purely voluntary and not binding. Therefore, EIGA or its members make no guarantee of the results and assume no liability or responsibility in connection with the reference to or use of information or suggestions contained in EIGA's publications. EIGA has no control whatsoever as regards, performance or non performance, misinterpretation, proper or improper use of any information or suggestions contained in EIGA's publications by any person or entity (including EIGA members) and EIGA expressly disclaims any liability in connection thereto. EIGA's publications are subject to periodic review and users are cautioned to obtain the latest edition. EIGA 2014 - EIGA grants permission to reproduce this publication provided the Association is acknowledged as the source EUROPEAN INDUSTRIAL GASES ASSOCIATION AISBL Avenue des Arts 3-5 B 1210 Brussels Tel +32 2 217 70 98 Fax +32 2 219 85 14 E-mail: info@eiga.eu Internet: www.eiga.eu

Table of Contents 1 Introduction... 1 2 Scope and purpose... 1 3 Definition... 1 3.1 Publication terminology... 1 3.2 Activated carbon... 1 4 Cryogenic gas purifier incidents... 1 4.1 Helium purifier explosion... 1 4.2 Neon purifier explosion... 2 4.3 Helium purifier explosion... 2 4.4 Hydrogen purifier explosion... 2 4.5 Pressure relief protection... 2 5 Potential hazards... 3 5.1 Ignition mechanisms... 3 5.2 Condensed oxygen (or dilute oxygen) mixtures... 3 6 Safeguards... 3 6.1 Adsorbent selection... 3 6.2 Oxygen reduction... 4 6.3 Oxygen analysis... 4 6.4 Flow/pressure changes... 4 6.5 Regeneration temperature... 4 6.6 Purging... 4 6.7 Regeneration pressure... 4 6.8 Operating conditions... 5 7 References... 5 Amendments to 43/07 Section Change Editorial to align with EIGA style manual 3.1 New section Publication terminology 5.2 Condensed oxygen (or dilute oxygen) mixtures section rewritten to clarify the section and add more information 6.2 Reinforcing the requirements for design 6.3 The requirements for oxygen analysers clarified and the need for a risk assessment. 6.7 Further advice on regeneration pressure NOTE Technical changes from the previous edition are underlined

1 Introduction As of 2013, there were three incidents reported to IGC and SAC of explosions in cryogenic activated carbon purifiers cooled by liquid nitrogen and one in an activated carbon cold adsorber in the nitrogen cold box section of the hydrogen liquefaction process. Each purifier had previously operated safely for many years. In all four cases it is believed that condensation of an oxygen enriched phase occurred and this reacted with the activated carbon adsorbent, resulting in an explosion. 2 Scope and purpose The document provides guidance on the safe operation of cryogenic activated carbon gas purifiers with particular reference on how to prevent explosions. 3 Definition 3.1 Publication terminology 3.1.1 Shall Used only when procedure is mandatory. Used wherever criterion for conformance to specific recommendation allows no deviation. 3.1.2 Should Used only when a procedure is recommended. 3.1.3 May and Need Not Used only when procedure is optional. 3.1.4 Will Used only to indicate the future, not a degree of requirement. 3.1.5 Can Indicates a possibility or ability. 3.2 Activated carbon Activated carbon, also called activated charcoal, is a general term which covers carbon material mostly derived from charcoal. It is a material with an exceptionally high surface area. Activated carbon has the property of adsorbing large quantities of gases. 4 Cryogenic gas purifier incidents Details of the four incidents are as follows: 4.1 Helium purifier explosion An activated carbon bed helium purifier exploded whilst purifying helium containing 1.4% oxygen and 0.7% nitrogen at 200 bar. Parts of the vessel were recovered up to 250 metres away. The explosion occurred when the operator opened a manifold valve to charge a second batch of cylinders. 1

It is believed that a liquid phase containing approximately 85% oxygen had formed and soaked into the activated carbon which was ignited by the flow and pressure surge when the helium was switched from the full to the empty manifold. 4.2 Neon purifier explosion An explosion occurred in an activated carbon trap of a neon purification train. The explosion, which occurred during heating for regeneration of the activated carbon, fractured both the cylindrical pressure vessel holding the activated carbon and the surrounding atmospheric pressure vessel for holding liquid nitrogen. The gas stream passing through the cryogenic activated carbon purifier contained 3% to 4% oxygen in a neon/helium mixture with approximately 1% to 2% nitrogen present. It is believed that oxygen rich liquid accumulated in the activated carbon, which ignited during the regeneration phase (carried out by electrical heaters thought to be capable of creating hot spots with temperatures of some 500 C). 4.3 Helium purifier explosion A laboratory helium purifier was being regenerated. The purifier had been operating for between two to three days, processing helium containing approximately 5% air. Prior to regeneration, the system was depressurized and liquid was drained off from the condenser. The purifier valves were closed. The explosion occurred whilst the purifier was being heated electrically. Significant damage occurred to the building structure. Examination of the stainless steel purifier fragments after the explosion revealed evidence of excessive high temperatures, probably due to internal heating from an activated carbon/oxygen reaction prior to the explosion. The ignition source is unknown but ignition could have occurred as a result of a pressure failure of the vessel due to corrosion. 4.4 Hydrogen purifier explosion An explosion occurred in a hydrogen liquefier commissioned in 1987. The main explosion took place in the nitrogen cold box section of the liquefaction process in an activated carbon cold adsorber vessel. The force of the explosion was assessed by using a mapping of the debris and has been estimated between 10 to 100 kg TNT equivalent. Process records showed that the main explosion occurred at the beginning of the regeneration phase of the activated carbon adsorber. When the explosions took place, the outlet temperature of the bed was still at 190 C. The catalyst used in the oxygen reduction system upstream of the hydrogen purifier was rapidly inhibited during operation, most likely by an unexpected contaminant in the feed from the unpurified hydrogen source. As a result, oxygen passed downstream into the purification section and was adsorbed on the activated carbon of the cold adsorbers. The regeneration blower generated a spark, which ignited a hydrogen/oxygen mixture and triggered a flashback to the adsorber bed. Main explosion in the adsorber vessel was due to activated carbon reaction with the adsorbed oxygen. 4.5 Pressure relief protection A common feature of all four incidents was that the safety relief valves were found to be operational following the explosions. This indicates that the reaction is so rapid that the purifier vessel cannot be protected by conventional pressure relief devices. 2

5 Potential hazards 5.1 Ignition mechanisms The ignition mechanisms of these incidents are not fully understood, however it is known that activated carbon/activated charcoal in the presence of condensed oxygen or certain dilute liquid oxygen mixtures (e.g. liquid air) is an unstable combination, which can result in a powerful explosion (1kg oxygen adsorbed in activated carbon is 2.6 times more powerful than an equivalent weight of TNT), see references [1] and [2]. Possible ignition sources include: high temperature or hot spot on regeneration (Auto-ignition temperatures as low as 200 0 C have been measured for activated carbon adsorbent in gaseous oxygen); the impact of particles entrained in the gas stream; rapid flow change through the bed (e.g. during operation of process flow valves), and heat generated by adiabatic compression. NOTE The presence of other flammable impurities (e.g. oil carryover from compressors or hydrocarbon contamination of the gas) is an additional hazard and should be removed before cryogenic purification. 5.2 Condensed oxygen (or dilute oxygen) mixtures As noted in section 5.1, the concern is the formation of a condensed liquid phase, containing either oxygen or a dilute oxygen mixture. This condensed phase could potentially react with activated carbon/activated charcoal. Such mixtures can potentially be formed by one or more of the following mechanisms: the feed to the adsorber contains a condensed liquid phase e.g. by bulk overflow, entrainment etc. from an upstream pre-condenser system; the feed to the adsorber contains excessive amounts of gaseous oxygen, such that a liquid phase will condense at the adsorber operating conditions. This can be caused by failure of upstream oxygen removal process e.g. by contaminants or saturation, or by the accidental ingress of oxygen during upset conditions in the upstream process (e.g. vacuum or low pressure process); upset during regeneration producing excess desorbed oxygen. 6 Safeguards If the process design incorporates a cryogenic activated carbon adsorber, a risk assessment should be performed to identify the potential ignition sources and the potential for producing a flammable carbon-oxygen mixture. The siting of this equipment should be carefully considered, and the risk should be minimized (although it cannot be eliminated) by taking the following precautions: 6.1 Adsorbent selection In some cases, acceptable performance can be obtained from alternative adsorbents, which are noncombustible (such as silica gel or molecular sieve), and thus inherently safe. Molecular sieve 5A has been shown to be nearly as effective as activated carbon in certain adsorption processes (reference [3]). When the adsorber feed can contain high oxygen concentrations, such alternative adsorbents should be considered. 3

6.2 Oxygen reduction If the feed can contain high oxygen concentration, then an oxygen reduction system shall be designed and be installed prior to the absorber, for example: a pre-condenser e.g. liquid nitrogen trap and knockout pot system, or a catalytic reactor. 6.3 Oxygen analysis When the plant contains an oxygen reduction system, then the oxygen content of the process gas shall be continuously monitored downstream of the oxygen reduction system. The high oxygen can occur by a process upset or by contaminants poisoning the catalyst. If there is no oxygen reduction system but the feed to the adsorber system can contain high oxygen in an upset case then the feed shall be analysed for high oxygen. If the regeneration gas is used in a recirculation mode there can be the potential to introduce oxygen and if this is the case an oxygen analyser shall be installed in the regeneration gas line. The adsorber system shall be shutdown if the oxygen concentration exceeds the maximum design concentration. The alarm set point should be provided by the manufacturer. If no set point is provided; then it shall be no higher than 50% of the maximum oxygen concentration for which the cryogenic purifier is designed, see reference [4]. The number, position and safety interlocks of oxygen analyzers should be defined on the result of a risk assessment. 6.4 Flow/pressure changes Sudden flow or pressure changes through the activated carbon absorber should be avoided by design and operating procedure. Apart from increasing the dusting of the activated carbon, it increases the likelihood of ignition. Particular attention should be given to cylinder bank change over and depressurising / repressurising of the adsorber. 6.5 Regeneration temperature Safety protection systems should be designed to ensure that the activated carbon will not exceed 100 C during regeneration, unless specific oxygen compatibility tests have proved it can be safely operated at higher temperatures. The temperature should also be slowly raised over a period of time to prevent local heating. 6.6 Purging Purging the bed with an inert gas will reduce the risk of ignition if high oxygen concentrations are present. Therefore, when possible, purge the activated carbon adsorber with an inert dry gas. Alternatively, if vacuum is pulled on the adsorber, any oxygen will be removed from the system before the adsorber is put back online. Purging at room temperature may not be sufficient to remove all the oxygen. 6.7 Regeneration pressure Cryogenic purifiers shall be regenerated at atmospheric pressure. If the purifier is blocked in during heating, the pressure rise due to desorption of contaminants and vaporization of any liquid phase might put a demand on the adsorber pressure relief device and if this failed could cause rupture of the vessel. The practice of closing valves to observe a pressure rise to measure regeneration performance is unsafe and shall not be used. 4

Where regeneration is achieved by applying reduced pressure (i.e. below atmospheric pressure) it should be noted that the removal of heat by convection of the gas is reduced and the risk of air in leak increased and therefore it is recommended that reduced pressure regeneration is restricted to the final regeneration cycle of the process. 6.8 Operating conditions Years of satisfactory service without incident should not be taken as proof of safe operation of cryogenic activated carbon adsorbers The design or process parameters of the purifiers shall not be changed without following a Management of Change process, see EIGA Doc 51, Management of Change [5]. All personnel operating activated carbon cryogenic gas purifiers shall be trained and qualified. 7 References [1] W.E. Tournay, F.M. Bower, F.W. Brown. "Safety and Performance Characteristics of Liquid Oxygen Explosives", US Bureau of Mines Bulletin 472 (1949). [2] G.S.T.J. Perrott, N.A. Tolch. "Liquid Oxygen Explosives", US Bureau of Mines Bulletin 349 (1932). [3] R Scott Willms Cryogenic Adsorption of Low-Concentration Hydrogen on charcoal, 5A Molecular Sieve, Sodalite, ZSM-5 and Wessalith Day, IEEE/NPSS 15 th Symposium on Fusion Engineering, October 11-15,1993 Hyannis, MA, USA. [4] Herring R.N., Barrick, P.L. "Gas-Liquid Equilibrium Solubilities for the Helium-Oxygen Systems". Adv. Cryogenic Eng., 10, pp 151-159 (1964). [5] EIGA Doc 51 Management of Change. 5