Final Report Summary - SAFERUBBER (A Safer Alternative Replacement for Thiourea Based Accelerators in the Production Process of Chloroprene Rubber)
Executive Summary:
A review of published work was undertaken to compare the already-suggested mechanisms and to further guide subsequent experimental work. In order to design better curing agents for chloroprene rubber, studies of the curing mechanism using advanced analytical techniques and computational methods continued. It has been found that, as expected, the mechanism of curing changes depending on the chemical used as the crosslinking agent. Our results lead us to the conclusion therefore that the mode of action of ETU cannot simply be described as one mechanism, but rather is a combination of multiple crosslinking reactions which change depending on the reaction conditions employed.
From the desk-based review of toxicity the key message is that ETU, the present used accelerator, has been classified as Reproductive toxicant category 1B with Hazard statement H360D – may damage the unborn child. Being classified as CMR, ETU is a substance that could be identified under REACH as a Substance of Very High Concern (SVHC) and therefor a potential candidate to be included in the Candidate List and later in Annex XIV of REACH (Authorisation).
The REACH Regulation mentions (Q)SARs as a mean to provide data without doing animal testing. ECHA (European Chemical Agency) presented a paper in 2009 informing that results of QSAR may be used instead of testing when certain criteria are met. As part of Annex VII testing for REACH registration (volume 1-10 t/y) and also Article 26 notification to ECHA is important. This work is currently on-going.
A life cycle assessment was carried out using the 4 steps: Goal and scope definition, inventory Analysis, Impact Assessment and Interpretation. Also tests to reduce/replace Zinc Oxide in the compound were carried out as well as leaching trials. Some findings from these tests relevant for environmental impact are
• the ability to flow and fill easily the mould has been found to be much better in the CR rubbers made with SRM102 than in the standard ones. This could arise in a reduction of the mass of rubber needed to vulcanize a piece which in turn will reduce the scrap.
• The cure cycle could easily be reduced through optimisation of the accelerator system
• It is possible to reduce the level of ZnO, by using either TiO2 or MFA’s especially when SRM102 is used as the accelerator.
• After 12 weeks leaching samples look similar. However, algae growth on samples produced by use of SRM102 suggests they’re not toxic!
Industrial validation trials were carried with chloroprene compounds containing SRM 102 and were tested in a wide array of industrial processing techniques in five factories in three different countries. In every case SRM102 has performed as well as, or better than compound containing ETU. It has proved itself to be a viable replacement for ETU as an accelerator. The rubber manufacturing trials have proven that the SRM102 is also a commercially feasible replacement for ETU. All industrial validation trials carried out during the project indicate that the proportion of SRM102 used in the formulas for all the end users will be very similar to ETU i.e. around 1%, therefore the current price estimates of the accelerator SRM102 are not anticipated to increase the manufacturing cost of a final rubber compound more than 105% of the equivalent compound made using ETU.
Dissemination has been active throughout the project and high visibility achieved for project partners and project results resulting in: More than 41 press quotes, 9 articles in rubber technology & science magazines (1 peer reviewed) + 1 master thesis dissertation, 4,890 unique visitors of the project website until May 2013. Dissemination of information by consortium members at 61 EU and global rubber industry events (conferences & trade fairs): 832 participants at conferences and a potential of 700,000 trade fair visitors. Presentation of project results at 5 Rubber Technology and 2 Chemometrics seminars in UK, Belgium, France and Italy. Training achieved of all project partners on the dissemination of the project results by the interactive development of a Dissemination Presentation Tool
Project Context and Objectives:
The primary objective of the SafeRubber project is to develop a commercially viable replacement accelerator molecule that can effectively replace thiourea based accelerators and metal oxides in the vulcanisation of rubbers such as chloroprene rubber. This is to be accomplished through the development of a new multifunctional additive and will be developed for use in existing processing equipment with subsequent development work taking into account processing conditions to provide an accurate grading of the resultant product, in terms of physical and chemical properties.
In order to achieve this overall aim, the consortium need to overcome a number of technical hurdles, which have defined a list of key objectives listed below:
1.To develop a replacement accelerator molecule (based upon new molecule or modification of existing molecule) that can produce polychloroprene rubber that performs to within ±10% of the parameters obtained with existing chemistries (thiourea based accelerators) such that the final product must:
a.Cure at a temperature and time similar to that obtained with the ETU/CR process (comparable Mooney Scorch curve at 130°C in accordance with ASTM D1646 or equivalent standard)
b.Enable materials to be produced with a tensile strength in the region 15-25 MPa, compression set of 15% after 24hrs at 70°C, modulus at 100% of 5.6 MPa and hardness of 74 IRHD (ISO 48:2007 or equivalent standard)
c.Have an operating temperature range of -45 to 120°C
d.Have excellent heat, oil and chemical resistance (in accordance with ISO 2475 or equivalent standard)
e.Not scorch or discolour during rubber processing (in accordance with ASTM D2084 or equivalent standard)
f.Have good UV, ozone and weathering resistance (in accordance with ASTM D0750-06/D1148-07 or equivalent standard)
g.Have an odour during processing and use that is no worse than ETU, as assessed by RBL’s Odour Panel.
2.To develop a safer alternative accelerator to ETU that complies with REACH and European Health and Safety requirements and is commercially viable at less than 105% of the equivalent product made using ETU.
3.To develop a scalable process for the large-scale production (~200 tonnes p.a.) of the replacement accelerator molecule such that the manufacturing cost of a final rubber compound is less than 105% of the equivalent compound made using ETU.
4.To demonstrate a scalable process by producing a large quantity (1-10 tonnes) of replacement accelerator molecule for use with Industrial Scale Validation trials.
5.To conduct industrial scale validation trials to demonstrate the performance of the replacement accelerator curing system in comparison to existing products and formulations on a industrial scale.
In addition to these project objectives, there are two further valuable contributions which the project will achieve:
a)Enhance the understanding of the fundamental mechanism of action of the accelerator system, based upon Ethylene Thiourea (ETU) and MgO/ZnO, which will allow further advances in the general understanding of the curing of polychloroprene and associated rubbers.
b)Enhance the understanding of using QSAR methodology to provide information on the feasibility of potential molecules through understanding how the molecular structure will affect toxicity while maintaining performance properties, especially performance properties especially as the current system is not based on an API (active pharmaceutical ingredient) for which QSAR is traditionally used.
Project Results:
Development of New Active Accelerator Molecule
Mechanistic Study
In order to design a suitable ETU replacement in the Saferubber project, the precise mode of action of ETU must be understood. To begin to understand this, a review of published work was undertaken to compare the already-suggested mechanisms and to further guide subsequent experimental work.
The curing of chloroprene monomers was next investigated. A significant delay was experienced at this point as the chloroprene monomer was not commercially available and had to be synthesised in the laboratory. The monomer was found to be quite unstable, undergoing gradual self-polymerisation over a number of days. In addition, cross-linking experiments with ETU could not be completed as the low boiling point of the monomer meant polymerisation and/or decomposition occurred before reaction with ETU was observed.
In parallel with Task 1.3 initial investigations into the curing of low molecular weight oligomers of chloroprene were also carried out to gain an understanding of the reactivity of the material. The initial results from this experimental work using FTIR and NMR analyses as suggested in the Description of Work did not yield a definitive answer to the mode of action of ETU as a wide spectrum of chloroprene mixtures was not prepared. However, the experiments have provided enough information to permit appropriate experimental design in the subsequent tasks and to suggest alternative analytical techniques for Tasks 1.2 and 1.3.
In order to design better curing agents for chloroprene rubber, we have continued to study the curing mechanism using advanced analytical techniques and computational methods. Using both chloroprene oligomers and gumstocks, Fourier Transfor infra-red (FTIR) spectroscopy and gas chromatography-mass spectrometry (GC-MS) have been employed to probe structural changes in the compounds to try to identify what is happening during crosslinking.
Kinetic studies show that zinc oxide (ZnO) rearranges the 1,2-isomer before crosslinking takes place. FTIR spectroscopy was used to monitor the disappearance of the band at 925 cm-1 (attributed to the C=C group in the 1,2-isomer) during crosslinking without additives and with ethylene thiourea (ETU), ZnO, and both ETU and ZnO together. When polychloroprene gumstock was heated to 160°C (without additives) there was a zero-order depletion of the 1,2-isomer during the first 25 minutes, reduced by over 90%. The remaining isomer was removed more slowly in the subsequent 10 minutes. This shows that the rearrangement of the 1,2-isomer can occur slowly on heating without any additives. With ETU as the only additive, the 1,2-isomer reduced much quicker. Approximately 60% was removed within the first 60 seconds. The remainder of the 1,2-isomer steadily depleted over the final 25 minutes. This confirms that ETU rearranges the 1,2-isomer quicker than when no additives are used at all. When ZnO is used as the only additive, the 1,2-isomer rearranges faster than under any of the other tested conditions, completely disappearing within a minute. With both ZnO and ETU as additives, the rearrangement is fast. There is a 90% reduction of the 1,2-isomer within one minute and the remainder disappears within 10 minutes. This is slower than with ZnO alone and suggests that there is a strong interaction (or complexation) between ZnO and ETU, which competes with the reaction of ZnO with the 1,2-isomer.
Systematic studies on the crosslinking of polychloroprene confirm that there are a number of competing mechanisms occurring at varying rates, rather than one single, straightforward reaction. The various reactions that have been identified include; (i) polychloroprene cross-linking with ETU through both its amine groups, as postulated by Kovacic (1955), (ii) cross-linking with ETU through one amine group and one sulphur, in the presence of ZnO and (iii) cross-linking of sulphur from ETU (in the presence of ZnO), which sees the formation of ethylene urea (EU) as a by-product. A possible route for the latter mechanism was published by Pariser (1960). Our studies have revealed evidence for each of these, or similar, competing mechanisms. When ETU and ZnO are both used, EU is evolved as a by-product, as confirmed by FT-IR and GC-MS, which is strong evidence for the Pariser mechanism. However, the published mechanism indicates that ETU reacts first with polychloroprene, before ZnO. However, we have strong evidence against this, as our kinetic study shows that ZnO reacts significantly faster with polychloroprene. Our computational analysis work also disagrees with the published mechanism and supports our own proposed mechanism.
Based on our current understanding, diamines such as 1,4 diaminobutane in combination with tetramethyl thiuram disulphide (TMTD) (as a less toxic model for ETU) were trialled as potential candidate accelerators. As predicted, this new accelerator combination was successful in crosslinking polychloroprene. The presence of the diamine appears to activate the allylic position of the 1,2-isomer on the polychloroprene chain in a similar manner to ZnO, albeit via a different mechanism, through which TMTD is then able to cross-link. It is possible that HCl released becomes incorporated as the diamine HCl salt. It should be noted that the diamines are also able to crosslink polychloroprene alone albeit poorly, as observed from results obtained through rheology and mechanical testing. The above observations give further confirmation of our understanding of the mechanisms occurring and allowed us to design new, safer accelerator salts containing anionic sulphur centres and cationic diamines that have been shown to be highly active for the curing of chloroprene.
Based on our current understanding, diamines such as 1,4 diaminobutane in combination with tetramethyl thiuram disulfide (TMTD) (as a less toxic model for ETU) were trialled as potential candidate accelerators. As predicted, this new accelerator combination was successful in crosslinking polychloroprene. The above observations give further confirmation of our understanding of the mechanisms occurring and allowed us to design new, safer accelerator salts containing anionic sulphur centres and cationic diamines that have been shown to be highly active for the curing of chloroprene.
During the final period of the project, work has continued in WP1 to understand the mechanism of curing of chloroprene rubber. This has built on the knowledge already gained on the mode of action of ETU from the previous periods of the project. The in-depth analysis of FTIR spectroscopy traces obtained from cured and non-cured oligomers and gumstocks has continued in order to identify the salient peaks and what they are caused by. Although the FTIR data had given us a deep understanding of the curing mechanism, the absolute identity of the peaks in the 1540 cm-1 region remained unknown. These are critical to understand, as they are attributed to the actual chemical crosslink in the polymer. Using chemical mimics for ETU, much work has been undertaken to observe the effect of the curing conditions on the peak sizes and positions in FTIR spectra. Tensile and other rheological data has then been obtained for the compounds cured with the chemical mimics to observe the effects caused by the replacement of ETU. Elemental analysis has been used as a probe for the chemical composition of cured products.
It has been found that, as expected, the mechanism of curing changes depending on the chemical used as the crosslinking agent. However, it has also been noted that proposed mechanisms of curing do not act independently and the FTIR data and tensile testing results suggest multiple cross-linking reactions take place, depending on the curing agent used; when zinc oxide and thioureas are used, bisalkylation dominates cross-linking; when thioureas or similar compounds are used, sulphur cross-linking dominates; when bases are used, cationic crosslinking dominates. Our results lead us to the conclusion therefore that the mode of action of ETU cannot simply be described as one mechanism, but rather is a combination of multiple crosslinking reactions which change depending on the reaction conditions employed.
To finally identify the peaks observed by FTIR spectroscopy in the 1540 cm-1 region, it has been proposed that Raman spectroscopy should be used in place of FTIR spectroscopy. It is believed that Raman spectroscopy may be able to distinguish between carbon-nitrogen and carbon-sulphur bonds, which would allow us to identify if bisalkylation or sulphur-crosslinking dominates in a particular reaction. However, Raman spectroscopy is a specialised technique and not widely available. A collaboration is being set up with a supplier of the technique and this work will continue outside of the project.
Confirmation of mechanism using gumstock resin
Gumstock resins were studied in this task in order to compare analytical results to the data obtained from the low molecular weight oligomers and the literature data. A set of gumstock samples was prepared containing various combinations of ETU, zinc oxide and magnesium oxide and cured under standard conditions. These were: gumstock and ETU; gumstock and ETU, MgO and ZnO; gumstock, MgO and ZnO; gumstock, ZnCl2 and ETU; gumstock, ZnO and EU (2-imidazolidone, ethyleneurea); gumstock and EU; gumstock and ZnCl2. Structural variations in the different uncured and cured rubbers were then investigated using solid-state FTIR spectrophotometry.
However, these data were not conclusive as only very minor variations in the FTIR traces were observed for each different sample; this was not unexpected as the structural changes are subtle and the curing agent is only present in very low concentrations. A wide range of other analytical techniques was therefore undertaken on the curing and cured gumstocks in order to support or reject suggested curing mechanisms and supply additional data.
In addition, computational analysis has been used to help validate and/or contradict the literature curing mechanisms. This analysis uses computational chemistry software on a supercomputer to predict likely transition states and intermediates during chemical reactions. The overall conclusion from this preliminary analysis was that the suggested mechanisms for the curing of chloroprene in the literature were possible but contained a number of unlikely transition states and intermediates. The information obtained from these computational analyses was combined with the experimental data to help explain observations and lend further weight to predicted curing mechanisms.
Molecular design of replacement accelerator and lab scale trials
QSAR modelling of replacement accelerator molecules
A thorough literature review was carried out to understand how and the extent to which QSAR has been used to estimate accelerator properties. It concluded that so far the use of QSAR to this end has not been exploited. A comprehensive bibliographic collection, which currently comprises 105 scientific papers related to different aspects of the world of rubbers, was generated. We retrieved from literature information about the characteristics of the vulcanization process itself (such as scorch time, optimum cure time, cure rate index, cure temperature, maximum and minimum torque) and the mechanical properties of the vulcanized (tear and tensile strength, elongation at break, hardness, crosslink density, modulus %, swelling index).
Documentation on available experimental data for existing accelerators: mechanical and rheological properties of rubber samples collected from the scientific literature were organised into a database. It includes 950 rubber samples (with different types of accelerator, formulation, rubber and temperature of vulcanisation), 11 mechanical and rheological properties measured on each sample, 5 qualitative variables defining type of accelerator, formulation, rubber, bibliographic source and vulcanisation temperature.
Statistical analysis of the mechanical and rheological properties of different compounds: the mechanical and rheological properties of different compounds were analysed by means of multivariate statistical tools (Principal Component Analysis) in order to understand the role played by different factors (type of accelerator, type of formulation, type of filler, type of rubber) in defining the rubber properties. A pattern mainly related to the bibliographic source was discovered. This suggested that a) experimental conditions/procedures really affect the measure of mechanical and rheological properties of compounds; b) the inter-laboratory variance seems to be the major source of discrimination of compounds. For this reason, a set of accelerators to be tested for their performance in vulcanization has been selected by RTD partners to obtain reliable data for modelling.
QSAR studies for prediction of curing properties: QSAR models attempt to relate structural features of accelerators (encoded by molecular descriptors) to their curing properties in order to predict properties of untested accelerators and select the best accelerator candidates. In order to set up the QSAR analysis, first the properties to be modelled were defined; these are optimum cure time, scorch time and cure efficiency. Then, a large set of 1650 molecular descriptors expected to be related to the curing properties was selected. Algorithms for the calculation of these molecular descriptors from the chemical structure files were generated. The most suitable multivariate statistical strategies to select the optimal subset of molecular descriptors for curing property calculation were evaluated. MATLAB codes to implement the above mentioned strategies and those for the calculation of mathematical models were generated. Finally, the molecular structures of the existing accelerators under analysis were designed and optimized by specific chemical drawing applications.
QSAR studies for prediction of toxicological end-points: existing QSAR models that relate chemical structures of substances to toxicological properties were evaluated and their efficiency was compared. The most effective models will be used to predict toxicological properties of accelerator candidates. Activities were initially focussed on a thorough review of documents on toxicological end-points relevant to REACH, along with QSAR scientific papers, shared by NORNER. Then, we carried out a screening of existing QSAR applications (i.e. CAESAR, TOXTREE, LAZAR, EPISUITE, QSAR TOOLBOX, SPARC, PBT PROFILER, VCCLAB, ONCOLOGIC), we evaluated toxicological end-points provided by each, the reliability of the predicted toxicological values, as well as their applicability domain. We mainly focussed on persistence, bioaccumulation, toxicity, toxicological, ecotoxicological and physicochemical properties, as suggested by NORNER and recommended in REACH annexes XIII and VII, respectively.
Existing QSAR models that relate chemical structures of substances to toxicological properties were evaluated and their efficiencies were compared. The most effective models were selected to predict toxicological properties of accelerator candidates. Activities were initially focussed on a thorough review of documents on toxicological properties relevant to REACH. The following properties were taken into considerations, since they were suggested by NORNER or recommended in REACH annexes XIII and VII, or commented in the SafeRubber project: bioaccumulation, biodegradability, carcinogenicity, skin sensitization, mutagenicity, aquatic acute toxicity, acute oral toxicity, developmental toxicity. A screening of existing QSAR tools (CAESAR, ToxTree, Lazar, Episuite, OECD QSAR Toolbox, SPARC, PBT profiler, VCCLAB, Oncologic, T.E.S.T) was carried out. Toxicological end-points provided by each tool, reliability of predicted toxicological values, as well as definition of applicability domain in each model were evaluated. CAESAR, ToxTree, OECD QSAR Toolbox, Lazar, Episuite and T.E.S.T. were selected as the most reliable QSAR tools and used for subsequent toxicological profiling.
ToxTree toxicological models are mainly based on a compilation of structural alerts and rules. These rules were analysed in order to enhance the selection of safe accelerators. Results of this analysis were shared with SafeRubber partners in November 2011. While analysing the ToxTree rules, some divergences between actual predictions and expected predictions were found. UNIMIB had a continuous contact with the ToxTree software webhouse in order to clarify these issues. It was found that some predictions provided by the Structural Alerts for the in-vivo micronucleus assay in rodents model according to rule “QSA34; H-acceptor-path3-H-acceptor” were erroneous; related predictions were thus manually changed and the bug was officially reported in the bug list of ToxTree. Moreover, the model for skin sensitization potential was modified in release v2.5.1 following discussion with ToxTree development team and detection of a bug on rule “Schiff base formation”. This rule includes several potentially sensitizing fragments, one of which being defined by the SMARTS number 12 [CH2][NH2]. In the developer’s intention, this SMARTS should detect only primary amines, but also secondary protonated (tetravalent) amines were erroneously detected.
Since toxicological screening dealt with several salts and some of the considered QSAR models could not give predictions for disconnected structures, corrective actions were undertaken. Initially, commercial QSAR softwares were evaluated (ACD/LABS, Simulations Plus, Accelrys), but none of them could support disconnected structures. Queries were submitted to the databases of OECD QSAR Toolbox in order to find disconnected structures for read-across implementation. Only the most similar compounds to the target salts were retained and this resulted in a total of 24 salts. This small set of disconnected structures was considered not appropriate to implement reliable read-across modelling. Finally, no publications dealing with QSAR models on disconnected structures were found in the scientific literature. As corrective action, it was decided to proceed with toxicological profiling of each ion present in disconnected structures, taken as separate entity.
QSAR models attempt to relate structural features of accelerators (encoded by molecular descriptors) to their curing properties in order to predict properties of untested accelerators and therefore provide information for the selection of the best accelerator candidates. In order to set up the QSAR analysis, first the properties to be modelled were defined; these were TS2 (scorch time), T90 (optimum cure time), minimum torque, maximum torque, the binary variable cured/not cured, tensile strength, modulus 100%, modulus 300%, elongation at break, hardness, as defined in the SafeRubber Annex I and in the project meetings of project Period 1. Then, a large set of 1650 molecular descriptors expected to be related to the curing properties was calculated. Algorithms for calculation of these molecular descriptors from chemical structure files were generated. The most suitable multivariate statistical strategies to select the optimal subset of molecular descriptors for curing property calculation were evaluated. MATLAB codes to implement the above mentioned strategies and those for the calculation of mathematical models were generated. Finally, the molecular structures of the candidate accelerators under analysis were designed and optimized by specific chemical drawing applications.
In order to get reliable QSAR models for curing properties of accelerators, experimental data for a set of commercial accelerators were provided by MaTRI. Ad-hoc QSAR models were calibrated on these experimental data. Due to linear correlation, the property Modulus 300% having large correlation with Modulus 100% was disregarded since equivalent QSAR models would have been found. For the same reason, cure efficiency, which was correlated to Maximum Torque, was excluded from the QSAR modelling.
Selection of replacement accelerator molecules
In order to produce the list of accelerator molecules to take forward to Task (Synthesis of accelerator molecules), it was necessary to define the chemical space that included potential accelerators and all the information needed to find a replacement accelerator molecule with similar performance to thiourea based accelerators, while avoiding their negative toxicological effects towards human health and environment. Initially, potential accelerator candidates were selected and included in an extended library on the basis of the following procedures: a) library design on chemical scaffolds suggested by the SafeRubber consortium and b) screening of “safe” molecules in toxicological databases through structural similarity analysis to existing accelerators. The mode of action of the accelerator system was not exactly defined in WP1 (Development of New Active Accelerator Molecule). Nevertheless, RBL from their experience suggested 12 new molecules and these were also chosen for analysis into their suitability as alternatives for ETU. Multifunctional additive materials, based upon bis-amino compounds containing a thione group (C=S bond) and an additional S–S linkage in the molecule, were selected for further study as potential new accelerators for the curing of chloroprene rubber. Upon heating, the multifunctional additive breaks down to yield a bifunctional amine and a thiuram disulfide species whereby it is believed the bifunctional amine will act as an accelerator for the vulcanisation process and the thiuram disulphide species will act as a cross linking agent.
Evaluation and definition of the criteria to select the final accelerator candidates: The following factors were defined to be relevant to the selection of replacement accelerators: safety/toxicity, curing and mechanical performance, potential economic cost of synthesis, potential feasibility of synthetic route. For each of these factors, a screening aimed to find the most specific indicators for accelerator selection was carried out. Indeed, for each toxicological property, a QSAR-based score was derived. Experimental values, if available, were considered instead of the QSAR predictions. Toxicological scores, mechanical, rheological, physicochemical properties and the indicators of economic cost and feasibility of synthetic route were evaluated for each molecule in analysis and finally combined to define a global accelerator score by means of Multicriteria Decision Making strategies. This global accelerator score accounted for all the relevant information of each molecule and allowed the molecules to be ranked from the most to the least preferable ones.
Prediction of toxicological and curing properties: All of the aforementioned indicators were evaluated for each molecule included in the extended library of potential accelerator candidates. This work was undertaken computationally by applying QSAR tools or through literature review in order to reduce time and costs of analysis and avoid animal testing. Molecules in the extended library were finally ranked on the basis of the selection criterion derived from the predicted indicators by means of Multicriteria Decision Making strategies. Only the first potential accelerator molecules were retained for further evaluation. Toxicological profiles based on QSAR predictions were also evaluated for all of the molecules proposed on the basis of experts’ knowledge about vulcanisation process in order to ensure toxicological safety of the final selected accelerators.
The final list of accelerators to take forward was selected considering both the molecules selected from the extended library and the molecules proposed by experts within the SafeRubber consortium. This selection comprised 15 molecules and was carried out by a panel of experts within the SafeRubber consortium. These molecules were numbered SRM100 (SafeRubber Molecule 100) upwards for identification purposes. The final 15 molecules were submitted to a deeper toxicological evaluation based on both QSAR models and a thorough search for experimental data by means of the available toxicological databases. More QSAR models were considered at this stage, since in the meanwhile several QSAR toolboxes released new models. Finally, the toxicological profile of ETU was compared with those of the 15 SRM accelerators. All 15 selected accelerators can be considered potentially safer than ETU, this being classed as carcinogen and mutagen on the basis of experimental results and classed toxic to reproduction in the CLP regulation
The majority of SRM molecules (14 out of 15 available molecules) are disconnected structures (salts). Only Lazar and ToxTree models can deal with predictions for salts. The toxicological properties which could not be predicted by means of QSAR models for SRM salts due to the lack of suitable QSAR models. Moreover, details, amendments and comments on the toxicological profiles of SRM accelerators were derived and shared with the project Partners. In particular, four molecules were predicted by the Structural Alerts for the in-vivo micronucleus assay in rodents model of ToxTree as being mutagenic according to rule “QSA34; H-acceptor-path3-H-acceptor”. The following considerations were drawn: a) this rule has a True Positive Rate (TPR, percentage of correct positive classification) equal to 34%. This is the lowest TPR in the decision rule of the ToxTree model and it derives that under this rule the considered QSAR model tends to classify as mutagenic many non-mutagenic molecules; b) the two considered Lazar models on mutagenicity gave for the same molecule non-mutagenic result; for these reasons, the four SRM molecules were considered as potentially non-mutagenic molecules. Regarding the ToxTree carcinogenicity model, one of the rules that identifies thiuram as carcinogenic molecule is a nitrogen atom (connectivity 3) bonded to a carbon atom, double bonded to S (connectivity 1) and to any other group but O or S bonded to C sp3, OH, SH, S- and O-. It derives that SRM salts are not identified as carcinogens, but the corresponding covalent molecule would be identified as carcinogen. This information was shared with the project Partners and further detailed in Deliverable 5QSAR methods were applied in order to analyse how molecular structure of Chloroprene rubber accelerators relates to their rheological and mechanical properties. QSAR models were developed in order to disclose which structural features mainly affect the mechanism of vulcanization. Regression mathematical models were developed for two rheological properties (Scorch time and Optimum cure time), and three mechanical properties (Modulus 100%, Hardness and Elongation at break). QSAR models were calibrated using experimental values of fourteen accelerators belonging to diverse chemical classes. A structural interpretation of each QSAR model was given, drawing hypotheses on the correlations between specific structural features and the analysed rheological and mechanical properties as well as defining possible chemical patterns connected to the mechanism of vulcanization.
Task 2.3 Synthesis of accelerator molecules
Nine of the twelve proposed molecules were synthesised as possible accelerators. A decision to leave two of the molecules out was taken as they were in the Class C category and would be included later if no suitable accelerators were identified in the Class A and class B molecules.
The synthesis of each molecule was carried out on a small laboratory scale up to 20 grammes and samples sent for curing trials.
Task 2.4 Lab-scale rubber compounding trials
The first phase of work involved testing many existing types of accelerator in CR compounds in order to build up a thorough knowledge of their effects on cure and to enable input into the QSAR modelling. Some of these have been used in CR previously and some have not. Both single accelerators and combinations of accelerators were tested. Additionally other chemicals, such as sulphur for instance, were used in some formulations to assess their effect on cure.
The second phase involved obtaining benchmark data for ETU so that comparisons can be made between it and the new SafeRubber molecule later in the Project.
Phase three consisted of testing all of the newly synthesised replacement molecules to assess their suitability for ETU replacement. At the conclusion of phase three, two molecules were identified as potential candidates suitable for up-scaling.
Laboratory scale industrial rubber processing and testing equipment was used throughout the trials.
All compounds were mixed using a 300 mm two-roll mill with a friction ratio of 1:1.5. The rolls were temperature controlled using the same setting of 20° C, front and rear, for all mixes. The CR was introduced to the machine on a tight nip and milled until a smooth band of rubber was formed on the front roll. The nip was then opened such that a full width rolling band was created in the nip to enable fast and efficient mixing. Because the compound temperature was low during mixing, all ingredients were added at the same time. Once the ingredients were dispersed the compound was cut off the mill and placed in a water bath to cool to ambient temperature. The compound was not left in the water bath once it was cool as CR compounds are known to absorb water.
To accommodate different size batches, the ears (side plates or guides) can be adjusted to alter the effective width of the rolls. The minimum batch size that can be mixed is around 100g, with the maximum batch size approaching 500g. During the SafeRubber Project, the largest batch that was mixed was approximately 275g.
If a compound proved to be capable of cure, test sheets and hardness slabs were then moulded using a 300mm x 300mm, electrically heated, hydraulic, up-stroking press. Compression moulds were used, one to produce hardness test buttons 30mm diameter x 20mm high, and the other to produce test sheets 220mm x 220mm x 2mm. The sheets were used to provide dumbbell shapes for tensile and other testing.
After mixing, each compound was tested on a Monsanto 100S rheometer to obtain its cure characteristics. The 100S is an oscillating disc-type rheometer. The rheometer plots a graph of torque verses time for any given cure temperature. The full extent of cure and beyond can be recorded. For example reversion, the point at which the cured compound breaks down due to prolonged heating can be observed and measured. Labline software was used to record and interpret the results.
Unless it was known from previous experience that a higher temperature was required for a particular accelerator, all test compounds were rheometer cured at 160° C to represent a typical industrial cure temperature. The remainder were cured at 175° C. The angle of arc was set at 3° with a frequency of 1.7 Hz.
A Testometric FS300CT tensile testing machine was used to carry out a range of tensile related tests: tensile strength, modulus at 100 and 300%, and elongation at break.
Hardness was measured using a Wallace bench-mounted, vibrating, hardness tester or a Wallace handheld hardness meter. Both were calibrated and gave identical results.
To obtain the maximum value from the trials, it was essential that variables between batches were minimised wherever possible. To provide a consistent baseline material for the laboratory trials, a 20 kg batch of masterbatch was mixed in one lot by Clwyd Compounders, using its production equipment. Although each trial batch still had to have the different accelerators added to it individually, this method reduced the risk of deviation due to inconsistent weighing of the core ingredients and variations in mixing.
After addition of the trial accelerators, the compounds were rheometer tested to establish their cure characteristics. Those which cured were then moulded and their physical properties assessed; hardness, tensile strength, modulus at 100% and 300% (if applicable), and elongation at break.
An indication of cure efficiency was obtained by deducting the figure for the minimum rheometer torque from the maximum torque. A low figure showing little cure and a high figure a good state of cure. From examining the results of the trials and comparing them to well-known parameters such as those of ETU, it can be stated that a cure efficiency of over 78 shows a good state of cure.
The first phase of compounding trials consisted of conducting a series of trials to establish the effects of a range of accelerators on the cure characteristics of a gumstock CR compound. Some accelerators were used in the masterbatch alone and some with the addition of a second accelerator or other curative. Along with others, all of the main rubber accelerator groups were represented whether or not they are generally used in CR compounds.
The information gained was used during the initial design of the new SafeRubber molecules. The effects of different molecular structures on the curing of CR could be assessed and taken into account when drawing up a shortlist of potential candidates.
If a compound showed little or no cure, physical testing (tensile, modulus, elongation at break, hardness) was not performed. In these cases, further trials were completed with higher concentrations of accelerator to establish whether or not they made a difference.
To enable comparisons to be made between the new SafeRubber accelerator and ETU, a series of six compounds was mixed for use as a benchmark. 1phr of ETU was used in the benchmark compounds because this is representative of the quantity used in CR compounds throughout the rubber industry. Both powdered (100% active) and pre-dispersed (75% active) forms of ETU were used.
Synthesised accelerators were numbered SRM100 (SafeRubber Molecule 100) upwards for identification purposes. As with the compound numbers, the nomenclature carried no further implication.
Very small quantities of some of the accelerators were initially produced due to the complex nature of the synthesis and difficulties encountered with their production. In some cases only 5g were produced. With a minimum, mixable batch containing 100 g of CR, this only allowed one or two tests to establish the synthesised molecule’s potential effectiveness. Generally, 2.5phr of the SRMs was used for the first test batch as it was considered that this quantity would be enough to establish whether there was any effect on cure.
In cases where the rheometer results showed good cure efficiency, and there was enough accelerator available to produce a 200g batch, square sheets were moulded for physical testing to be performed.
SRM102 and SRM104 both showed reasonable cure characteristics, with SRM102 being the best, in that is cured quicker with a more distinct plateau than the slight marching modulus of SRM104. The rheometer trace for SRM102 is comparable to that of ETU. In fact at 2.5 phr it shows a quicker cure with a higher final torque and no marching modulus as displayed by the ETU compound. Further optimisation will take place in WP5.
Both SRM102 and SRM104 were chosen for scale up trials.
Environmental and Commercial Impact
Desk-based review of toxicity
The work focused on ETU as a reference for the new accelerator system under development.
Ethylene thiourea (ETU) has been used as accelerator in production of chloroprene rubber for decades. ETU is classified as CMR (Carcinogenic, Mutagenic or Toxic to Reproduction) and is thereby a substance that could identified as a Substance of Very High Concern and afterwards included in Annex XIV of REACH (Authorisation).
ETU is classified as CMR (Carcinogenic, Mutagenic or Toxic to Reproduction) and is thereby a substance that fulfils the criteria for being identified as Substance of Very High Concern (SVHC) and therefore as a future candidate to be included in Annex XIV of REACH (Authorisation). The main purpose with REACH is to ensure a high level of protection of human health and the environment.
Hazardous substances will be subject for Authorisation or Restrictions to ensure safe use.
Authorisation is a complex and heavy process. Relevant substances need to go through a process before they eventually are included in Annex XIV:
1. Identification of SVHC: Member States (MS) or ECHA on behalf of Commission include substance in Registry of Intentions
2. Submission of REACH Annex XV dossiers
3. Publication of Annex XV reports for comments
4. Development of Support Document (incl. responses to comments) by MS and ECHA
5. Agreement on identification by ECHA’s Member State Committee (if comments are received)
6. Inclusion in “Candidate list”
7. Immediate obligations following inclusion in the ‘candidate list’ for substances, mixtures or articles containing a “Candidate list” substance
The last substances were included in the “Candidate list” on 20.06.2011 and the list contains now 53 substances. The first 6 substances for Annex XIV was published 17.02.2011 and a corrigendum was issued the day after (Regulation (EC) No. 143/2011).
Article 67 in REACH states that “A substance on its own, in a mixture or in an article, for which Annex XVII contains a restriction shall not be manufactured, placed on the market or used unless it complies with the conditions of that restriction”. ETU use is currently restricted due to its classification as “toxic to reproduction category 1B”: Shall not be placed on the market, or used, as substances, as constituents of other substances, or, in mixtures, for supply to the general public when the individual concentration in the substance or mixture is equal to or greater than: either the relevant specific concentration limit specified in Part 3 of Annex VI to Regulation (EC) No 1272/2008, or the relevant concentration specified in Directive 1999/45/EC.
The key message is that ETU, the present used accelerator, has been classified as Reproductive toxicant category 1B with Hazard statement H360D – may damage the unborn child.
Being classified as CMR, ETU is a substance that could be identified as Substance of Very High Concern (SVHC) and a potential substance for inclusion in the Candidate List and later in Annex XIV of REACH (Authorisation).
Toxicity Testing
Understanding of the toxicological implications related to the introduction of new accelerator molecules is important. In line with the European regulations testing on animals should be avoided if possible, and toxicology databases can be a tool to be used to prevent unnecessary use of animal testing.
QSAR (quantitative structure activity relationship) has been an important study in the second year of the SafeRubber project. The outcome of a systematic QSAR study by UNIMIB based on 9 chemical scaffolds coupled with proposed molecules from rubber experts in the consortium was 15 molecules. Delivery Report 10 gave the toxicology testing results based on the use of QSAR tools. Twelve molecules are predicted to be safer than ETU.
Based on the on-going curing tests 2 molecules, SRM102 and SRM104 have given very promising results.
Both molecules have been predicted to be not mutagenic, not carcinogen and not skin sensitizer. However based on the laboratory trials SRM102 was chosen to be the best molecule for further development.
The selected sub-contractor LAB Research (Hungary) was acquired by CIT in June 2011 and the name was changed to CiToxLAB (Hungary). CiToxLAB has ca. 800 employees and has facilities in Hungary, France, Denmark and Canada.
QSAR (quantitative structure activity relationship) has been important in the SafeRubber project to screen for potential safe substances and to demonstrate that selected accelerator candidates are safer than ETU. The main purpose of QSAR was therefore not focused on REACH registration.
The outcome of a systematic QSAR study by UNIMIB was that 6 substances were predicted to be not mutagenic, not carcinogen and not skin sensitizer. Of these substances SRM 102 was selected based on its curing properties as accelerator for polychloroprene.
The REACH Regulation mentions (Q)SARs as a mean to provide data without doing animal testing. ECHA (European Chemical Agency) presented a paper in 2009 informing that results of QSAR may be used instead of testing when the following criteria are met:
1. Results are derived from a QSAR model whose scientific validity has been established
2. The substance falls within the application domain of the QSAR model
3. Results are adequate for the purpose of classification and labelling and/or risk assessment
4. Adequate and reliable documentation of the applied method is provided
5. In cases where there is uncertainty related to one or more information elements, QSAR results may still be used in the context of a Weight of Evidence approach
The main challenge with SRM 102 regarding QSAR is that this substance is a salt. Due to this the following actions were done:
• A screening of commercial tools able to handle QSAR predictions on disconnected structure was performed, but all the considered QSAR tools (ACDLAB - ACD/Tox Suite, Simulation plus, Accelrys - Discovery Studio TOPKAT) cannot support disconnected structures
• Queries on the databases of the OECD QSAR Toolbox were carried out in order to find disconnected structures to be used for read-across models. Few experimental data were available for salts and mainly for different endpoints, organisms and test conditions, making them not comparable and suitable to make predictions. Therefore, the set of disconnected structures in the QSAR toolbox was considered not significant in order to enhance reliable read-across models
• Relevant papers published in the scientific literature reporting QSAR models on disconnected structures were not found
In practice this means that our QSAR data should be used as supportive data for REACH registration. This means that the needed data in REACH Annex VII should be based on test data.
CiToxLAB, the selected sub-contractor for REACH testing, started testing end February 2013 based on a sample of SRM 102 made in Robinson Brothers Ltd. pilot plant. Tests are still ongoing.
The results so far have shown that SRM 102 is safe and the test data confirms QSAR predictions on toxicity (details are given in Delivery Report 14).
As part of Annex VII testing for REACH registration (volume 1-10 t/y) also Article 26 notification to ECHA is important. This work has also been sub-contracted to CiToxLAB.
Towards the end of the SafeRubber project ECHA announced in the “Registry of current SVHC intentions” that Sweden has filed a “notification of intention” of ETU (scope CMR) on the 16th April 2013. A likely consequence of this is that ETU will be included in the Candidate List during 2013.
QSAR validation
The entire set of ca. 52.000 molecules in the OECD QSAR Toolbox was considered initially by UNIMIB. By use of several “filters” the end result was the earlier mentioned 15 molecules.
A number of QSAR tools have been selected to get prediction of carcinogenicity, mutagenicity, developmental toxicity, acute toxicity, skin sensitisation, bioaccumulation and biodegradability.
QSAR tools designed to meet REACH requirements have been used. Such tools are OECD QSAR Toolbox, CAESAR, ToxTree, T.E.S.T. and Lazar.
Important criteria to meet when using QSAR are:
• Results are derived from a QSAR model whose scientific validity has been established
• The substance falls within the application domain of the QSAR model
• Results are adequate for the purpose of classification and/or risk assessment
• Adequate and reliable documentation of the applied method is provided
• In cases where there is uncertainty related to one or more information elements, QSAR results may still be used in the context of a “Weight of Evidence” approach.
The main purpose with QSAR in SafeRubber was to demonstrate that selected substances are safe. Validation and optimization of QSAR models have been an intrinsic part of the model construction in SafeRubber.
Environmental Impact Study
The initial plan was to make an evaluation of environmental impact of SRM102 throughout production, usage and disposal and compare by the impact of ETU, by use of ECETOC TRA (Targeted Risk Assessment) as offered by European Centre for Ecotoxicology and Toxicology of Chemicals (www.ecetoc.org). However, as the production of SRM102 has not been scaled up for commercial production yet and the toxicology testing by CiToxLAB has not been finalized the data are too few for a reliable TRA.
An ethic review was undertaken to minimize the use of animal testing by substituting QSAR results where applicable.
The scale-up was carried out at Robinson Brothers Ltd who are ISO14001 registered (Certificate No: EMS 57413) and before proceeding the established parameter check list was used as follows:
• Raw Material Assessment COSHH
• Safety Testing DSC
• Process Waste Analysis
• Quality Control Analytical Methods
• Energy Used
• Packaging Supply
• Disposal
• Equipment Required
• Transport Raw Materials
• Product to customer
• Odour
The following four steps have been defined for making a Life cycle assessment (LCA) according to ISO 14040 for replacement of ETU by SRM102 as accelerator in the production process of chloroprene rubber:
1. Goal and scope definition
• The objectives for doing the LCA comparison between SRM102 and ETU are:
• Better understanding of the value chain
• Insight on potential environmental impact risks
• Aligned economic, environmental and societal goals
• Engagement of value chain partners and stakeholders
• Identification of effective improvements
2. Inventory Analysis, including tracking material and energy flows in all relevant process steps:
This is an important part especially to collect the right information with respect to worker exposure and risk scenarios, and influence on the environment during accelerator production and vulcanization of chloroprene rubber, in addition to distribution, use and waste generation. Data needed to make an Impact assessment are:
• General physicochemical data of SRM102 and ETU (molecular weight, vapour pressure, water solubility, partition coefficients, in addition to results from biodegradability testing –
• In general, the routes of potential human exposure to SRM102 (and ETU) are inhalation, ingestion and dermal contact and the potential occupational exposure is greatest for workers involved in the manufacture of the accelerator and the manufacture of rubber and rubber products. Data needed for the worker exposure and risk scenarios (physical form, vapour pressure and operating temperatures. In addition information on ventilation, the use of respiratory and dermal protection to calculate inhalation exposure and dermal exposure are needed – (for SRM102 such data will be available when a scale up production has been designed, hopefully autumn 2013)
• Data on Energy inputs (track quantity, type and source of energy) – (data available after scaled up production of SRM102)
• Data on Production output including purity and eventually by-products – (data available after scaled up production of SRM102)
• Data on Emissions and Waste Outputs – (data available after scaled up production of SRM102)
3. Impact Assessment
For the Impact Assessment in this project it has been decided to use ECETOC TRA (Targeted Risk Assessment) as offered by European Centre for Ecotoxicology and Toxicology of Chemicals (www.ecetoc.org). However, as the production of SRM102 has not been scaled up for commercial production yet and the toxicology testing by CiToxLAB has not been finalized the data are too few for a reliable TRA.
An additional requirement was to try to reduce the heavy metal content (ZnO) of the compound and also to carryout leaching studies.
Reduction of heavy metal oxides
Polychloroprene rubber (CR) is commonly cross-linked by adding a combination of zinc oxide (5 parts per hundred rubber (phr)), magnesium oxide (4 phr) and an accelerator such as ethylene thiourea (ETU) or the new SafeRubber molecule, SRM102. Additional accelerators may be used to boost the cure rate.
According to the EPA “zinc ion can become available from zinc oxide through several mechanisms” and that “Zinc ion can reasonably be anticipated to be toxic to aquatic organisms”. One of the wider European benefits anticipated by the Safe Rubber Project is a reduction in the use of Zinc Oxide.
Two different experimental approaches were evaluated a) the replacement of ZnO by TiO2 and b) the replacement by a Multi-Functional Additive (MFA)
TiO2
A literature and patent search to identify previous work conducted with TiO2 was undertaken. It was identified that titanium dioxide has been used as filler in numerous elastomeric formulations but the use of this compound as a curing/vulcanising agent, in its own right (there are patents relating to titanium dioxide-containing complex organic molecules and titanium dioxide photocatalysts), is extremely limited (no patents were find to its use in polychloroprene formulations). However, a reference from this search was identified:
16/9/3. ‘Room temperature curable silicone rubber composition for use in e.g. electric appliance, composition comprises silicon polymer, surface treated silica, silane adhesive, silicone oil and titania type or zirconium-based curing catalyst’
Initial tests using TiO2 were performed at an early stage in the SafeRubber Project as part of a wider series of trials in conjunction with work investigating the curing mechanism of CR and also to provide data for QSAR computer modelling. When it was realised that TiO2, had the potential to replace zinc oxide (ZnO), it was decided to perform further work. The first series of tests were labelled SR134 to SR138. The five formulations were mixed and rheometer tested to establish their cure characteristics. The results are shown below. A measure of cure efficiency was established by deducting the minimum rheometer torque from the maximum.
It was established from this first series of tests that if ZnO is substituted by TiO2, when combined with the accelerators, HEXA and HVA enabled a good state of cure can be obtained. In fact the cure efficiency index is higher when TiO2 is used.
It was decided to investigate whether CR could be cured using TiO2 in place of ZnO using ETU as an accelerator with the addition of sulphur to the formulation and a good state of cure was obtained.
The optimum quantity of TiO2 was then investigated. Three compounds were formulated using 4 phr, 5 phr and 6 phr of TiO2, and these were based on the HEXA/HVA accelerator system.
Whilst the results of this brief test could be considered preliminary, it does appear that the TiO2 could be reduced to at least 4 phr without detrimental effect on the state of cure.
The final stage of work conducted during the Project was to investigate the effects of a reduction of sulphur and the combined metal oxides, in particular zinc oxide, or titanium dioxide, and magnesium oxide. Compounds SR245, SR246, SR247, SR248 and SR249 were used for this. The compounds were mixed and tested for their cure characteristics using a rheometer and for their physical properties.
The following conclusions can be observed from the results:
1. SRM102 gives a better cure than ETU, both in cure efficiency and physical properties.
2. A reduction of 1phr of each of the two metal oxides in the SRM cure compound does reduce the state of cure and physical properties; however the results are broadly in line with the ETU cured compound.
3. Reducing the sulphur content from 1 phr to 0.5 phr (active) at the same time as reducing the metal oxides does not greatly affect the cure.
4. If the sulphur is removed completely, reasonable cure efficiency is obtained but the physical properties drop considerably.
5. It is possible to replace zinc oxide with titanium dioxide in the curing system of polychloroprene rubber compounds.
6. It is possible to reduce the quantity of metal oxides from the current 5 phr ZnO and 4 phr MgO without greatly affecting the physical properties of the cured compound.
However there is still a great deal of further work which could be conducted, especially on the subject of ZnO replacement. The proportions of metal oxides and other components of the curing system all require optimisation and a full suite of physical testing needs to be performed once this is done, this work will continue post project.
Multi-Functional Additive (MFA)
A small amount of multi-functional additive (MFA) was synthesised to test whether it was possible to reduce the level of ZnO in a CR formulation. Mechanical tests have been completed with tensile testing results done after curing for 1.5 times T90, results including tensile modulus at various extensions, ultimate tensile strength and elongation at break were found. Similarly MDR cure results for a 15 minute test at 160 °C were found including maximum torque reached (MH), scorch time (TS1) and maximum cure rate.
The physical testing results show that the standard cure system produces similar results to SRM102 containing systems, in regard to both tensile results and cure results. It can also be seen that formulations 4 and 5 produce similar results to each other.
The conclusion from this small trial indicates that by incorporating MFA into a cure system it is possible to reduce the level of ZnO. It is recommended that partners who wish to use MFA carryout further optimisation work post project.
Leaching trials
In order to assess leaching of components of chloroprene rubber into rainwater over time, and thus assess environmental impact, a leaching experiment was set up. Samples of known weight (5g) of chloroprene rubber with different compositions were rinsed with rainwater and then placed into 25ml of rainwater and left for 12 weeks, with occasional stirring. This set-up aimed to mimic at a basic level the environment in a landfill or similar waste disposal site. After a 12-week period, the water samples were examined using UV-vis spectrophotometry to determine if any leaching of material had occurred.
The absorbance for all samples is very low. For the chloroprene gumstock sample cured with SRM102, no leaching of any components is observed after 12 weeks. For the other samples, absorbance values are all below 0.09 indicating exceptionally low concentrations of leached components. The lack of defined peaks also suggests the absorbance values observed are more likely to be down to light scattering by fines in the rainwater and machine error than due to absorbance by dissolved components.
Different samples of chloroprene rubber have been suspended in rainwater for 12 weeks and the water subsequently analysed by UV-visible spectrophotometry. Data obtained suggest that leaching of the component chemicals into the rainwater over the 12 week period is minimal and does not provide an environmental risk.
UV testing
UV testing was carried out according to EOTA (European Organisation for Technical Approvals) Technical Report specifies exposure procedures for artificial weathering using EN ISO 4892. The radiant exposure range in Northern Europe is typically 128MJ/m² per year.
The laboratory light source type 1 (UV-A 340 nm peak) was used where radiant emission below 400 nm makes up at least 80% of its total light output and where radiant emission below 300nm is less than 2% of its total light output, with a spectral irradiance of (45 ± 10) W/m2 in the bandpass of 300 nm to 400 nm.
The simulation exposure conditions were for a moderate climate in Europe. The testing cycle was UV radiation for 4 hrs at 60°C then a condensation cycle for 4 hrs at 50°C then repeated. The lamps run at ~40W/m² (300-400nm). Testing was for 500 hours, this is equivalent to half a year in the Arizona desert.
The test samples were black filled and gumstock moulded compounds:
• SR141, ETU gumstock
• SR179, SRM102 gumstock
Tensile testing before and after exposure was used to determine the effects of UV exposure and the results show that there is no difference between compounds made with ETU and SRM102 and both show that there is negligible degradation due to UV radiation.
Conclusion
Chloroprene compounds containing SRM 102 have been tested in a wide array of industrial processing techniques in five factories in three different countries. In every case SRM102 has equaled or exceeded compound containing ETU. It has proved itself to be a viable replacement for ETU as an accelerator Deliverable report D18).
4. Interpretation
The interpretation including limits, gaps, recommendations and conclusions will be made after the TRA has been finalized. However, Chloroprene compounds containing SRM 102 have been tested in a wide array of industrial processing techniques in five factories in three different countries. In every case SRM102 has equalled or exceeded compound containing ETU. It has proved itself to be a viable replacement for ETU as an accelerator Deliverable report D18). Some findings from these tests relevant for environmental impact are:
• Along the experience with the vulcanization trials, the ability to flow and fill easily the mould has been found to be much better in the CR rubbers made with SRM102 than in the standard ones. This could arise in a reduction of the mass of rubber needed to vulcanize a piece which in turn will reduce the scrap.
• The cure cycle could easily be reduced through optimisation of the accelerator system
• ZnO is very toxic to aquatic organism. It is possible to reduce the level of ZnO, by using either TiO2 or MFA’s especially when SRM102 is used as the accelerator; however further trials post project will be needed to investigate other effects such as ageing and oil resistance etc.
• Rubber samples have been placed into rainwater. After 12 weeks samples look similar. However, algae growth on samples produced by use of SRM102 suggests they’re not toxic!
• There is no difference in the UV degradability between ETU and SRM102 compounds.
Commercial Feasibility study
A commercial feasibility study will be undertaken throughout the project to ensure the development of the replacement molecule is commercially viable (is cost competitive to state-of-the-art materials i.e. the manufacturing cost of a final rubber compound is less than 105% of the equivalent compound made using ETU). The study will be directly relevant to the lab scale development in WPs 1&2 and will feed into the scale up activities in WP4 and industrial validation in WP5.
The trials carried out in WP 5.2 (rubber manufacturing trials) and reported in WP 5.3 have proven that the SRM102 is a commercially feasible replacement for ETU.
Robinson Brothers Ltd is currently unable to estimate the definitive price of SRM102 until it produces the product on an industrial scale (it has only produced 44 kg so far). However all of the trials carried out indicate that the proportion of ETU and SRM102 used in the formulas for all the end users is very similar i.e. around 1% , therefore the current cost estimates of the accelerator SRM102 will not increase the manufacturing cost of a final rubber compound more than 105% of the equivalent compound made using ETU.
For a more in-depth report showing tables and graphs, please refer to the attached PDF.
Deviations & Corrective Actions:
None
Work package 4: Scale Up of Replacement Accelerator
Development of suitable synthesis route and process conditions
According to the preliminary curing tests two possible new accelerators SRM102 and SRM104 were selected for Kilo Lab-scale process.
It was found that the syntheses of SRM102 and SRM104 accelerators were complicated and accompanied by the formation of several by-products. The formation of undesired products is due to similar pKa values and similar reactivity of both free amino groups in the starting compound. As a consequence a new, reproducible procedure for the preparation of the target products SRM102 and SRM104 on a 1 kg scale as one batch syntheses was developed. The reaction conditions, reagents, solvents, temperature and reaction time were been optimised.
All intermediates, by-products and final products were analysed by 1H NMR spectroscopy. The purity of the key intermediate and products SRM102 and SRM104 was analysed by HPLC-MS technique However, under standard conditions, HPLC-MS analysis is not very reliable. There is the possibility that the analytes can interact with sorbent or mobile phase. The greatest problem is pH of the eluent: exchange of eluent to pH>9 (suitable for solubility of betaine) will result in even lower sensitivity. The use of high pH values and various salt additives will result in extremely low lifetimes of HPLC standard columns.
The new, scalable and reproducible method for the preparation of the key intermediate betaine was elaborated. All possible alternative methods and by-products were examined and analysed by 1H-NMR. The optimal conditions were found and the intermediate betaine was obtained in 80-95% yield with the purity 99% by NMR.
With betaine as the key intermediate, the optimal reaction conditions of the target accelerators SRM102 and SRM104 have been determined and reported.
The purity of synthesised products: betaine, SRM102 and SRM104 was analysed using acid-base titration analysis. The acid-base titration is one method for the quality control of betaine and final products SRM102 and SRM104. In this case side interactions of target products and used materials for titration were not observed.
Scale up process on the Kilo Lab Scale
The new accelerator SRM102 was prepared using the synthetic procedure outlined in Deliverable 15 at Kilo Lab scale i.e. 1kg batch trial in quantitative (>99%) yield with the purity 97.5% combination of 1H-NMR (impurities) and acid-base titration (assay) and the new accelerator SRM104 was obtained in 1kg batch trial in quantitative (>99%) yield with the purity 99.1% combination of 1H-NMR (impurities) and acid-base titration (assay).
The bulk synthesis compounds were analysed and the results compared to the compounds synthesised at small scale using both 1H-NMR NMR and acid-base titration. Analysis showed the compounds to appear chemically identical.
To ensure curing efficiency was also maintained, bulk SRM102 and SRM104 samples were used in chloroprene curing experiments and the results compared to small-scale-synthesised SRM102 and SRM104 chloroprene cure curves. It is clear from the cure curves that the Kilo Lab produced samples of SRM102 and SRM104 behave in the same way as the small-scale-produced samples do and so scale up to pilot scale can now be evaluated.
Novel chloroprene curing agents SRM102 and SRM104 have been successfully produced at Kilo Lab scale and shown to behave in the same way as samples produced at small scale in the curing of chloroprene rubber.
Scale up process on Pilot Plant Scale
Kilo quantities of SRM102 and SRM104 have been made by GSL and a synthesis route determined and reported in D15. RBL are currently assessing the proposed synthesis routes. Further studies carried out by the research Partners on the two potential candidates has shown that SRM102 shows improved rheological and physico-mechanical properties over SRM104. As a consequence, the Consortium has agreed that scale-up should concentrate on SRM102 only.
A series of validation trials was carried out in the laboratory and then kilo scale in order to ascertain the robustness of the process and to obtain the critical Health and Safety data for the process before proceeding to Pilot Plant scale. At the validation stage, modification to the process may be required to improve and optimise the process and yield. It should be noted that process development and optimisation may alter the final product, e.g. purity, and this may have an effect on the efficacy of the accelerator molecule compared to the research developed/synthesised equivalent. This will need to be assessed by the research Partners.
Quantities of SRM102 manufactured at the validation stages will be made available to Clwyd Compounders and Mixer to allow them to carry out their own compounding optimisation trials whilst the pilot plant quantities are being manufactured.
RBL plan to carry out validation trials in June/July and scale up to Pilot Plant in mid-August/September. The time-scale provided assumes that the process provided will not require any changes. Any process development and optimisation will increase this time-frame. Validation trials on laboratory and kilo-scale will need to be carried out to determine the robustness of the process and to provide critical health and safety data required before the process can be carried forward to Pilot Plant scale. Any process improvements made to the process may alter the final product and therefore its efficacy may need to be reassessed.
RBL plan to start process validation trials in June and provided no process development is required, Pilot Plant scale-up is scheduled from mid-August/September. This time-frame will increase if process development and optimisation is required.
The aim of this subtask was to determine methods for the reproducible synthesis and scale-up of production of SRM102 and SRM104, going through kilo- and pilot-scale preparation. In order to achieve this, it was first necessary to determine a suitable synthesis route for scale-up, based on those outlined previously in WP2. Before the scale-up could be started it was necessary to ensure that the Robinson Brothers parameter check list was followed:
• Raw Material Assessment COSHH
• Safety Testing DSC
• Process Waste Analysis
• Quality Control Analytical Methods
• Energy Used
• Odours
• Packaging Supply
• Disposal
• Equipment Required
• Transport Raw Materials
• Product to customer
• Odour
During the lab-scale production of SRM102 and SRM104 the synthesis routes were modified slightly taking into account the above parameters and the synthesis was determined to be quite complicated and careful control was required in order to avoid making by-products. These synthesis routes were used as a starting point for the kilo-scale synthesis of the new accelerator molecules, employing 1H NMR spectroscopy and HPLC as analytical techniques to support synthesis optimisation.
It had already been observed in WP2 that successful production of the accelerator molecules depended heavily on the acidity of the reaction mixture, due to the varying pKa values of the reaction intermediates. If the acidity of the reaction mixtures was not carefully controlled, the reaction mixtures could decompose or undergo undesired reactions. By studying the various pKa values of the reaction intermediates and products, a synthesis route was determined that first produced a betaine intermediate, which was common to both SRM102 and SRM104. This betaine could then be further reacted with the desired diamine to produce the accelerator compounds.
With this knowledge, the synthesis route at kilo-scale was undertaken and optimised to yield SRM102 and SRM104 in 80 – 95% yields. The next important stage in the scale-up procedure is the development of a suitable assay method to determine purity. The most widely used purity assessment method is HPLC, due to its reproducibility and speed of analysis. This method was therefore selected for purity determination of these accelerator compounds. However, as SRM102 and SRM104 are salts, determination using HPLC is not straightforward. This is because salts dissociate when dissolved and so it is difficult to determine the purity of the compounds using one analysis method alone.
Initially HPLC-MS was used, but the results from this analysis method were not reliable or reproducible. It was thought that this may be due to the analytes interacting and reacting with the HPLC column material or decomposing during analysis. Another major issue was the need to run analyses at pH>9 to ensure solubility of the components. This leads to short column lifetimes and reduced sensitivity.
To overcome this, an acid-base titration technique was first determined to assess purity as a replacement for HPLC analysis. However, once again, the sensitivity of this technique was low and the variation in results was large. In addition, analysis times were too long and laborious to be used at pilot- and production-scale. Therefore, analysis by HPLC was pursued again, this time not employing mass spectrometry. After further optimisation of the analysis methods, it was decided to develop methods for the analysis of the anion and the cation separately. This was significantly easier than trying to devise a combined analysis method and the purity of the SRM102 and SRM104 compounds could therefore be assessed successfully.
The HPLC traces obtained for the SRM102 and SRM104 compounds showed that the individual dissociated anion and cation were both around 97% pure after the first kilo-scale synthesis by GSL. The synthesis and analysis methods were therefore handed over to RBL, where they were carried out at kilo-scale in their kilo-lab production facilities. Using HPLC and TGA analyses, the resulting products were compared to those produced by GSL and found to be similar, showing the synthesis procedure was reproducible and produced the desired products in high yield and purity. The compounds were also found to act as curing agents in the same way as the products produced at small scale.
Finally, the synthesis method was developed into a pilot-scale production method. It was decided on the basis of data on toxicology, safety and product quality that only the large-scale production of SRM102 would be undertaken. This was achieved at pilot-scale, with 44kg of SRM102 being produced and analysed.
Work package 5: Rubber Curing, Component Manufacture and Industrial Validation
Rubber Compounding Trials
The SRM102 production molecule was supplied to MGN, Mixer, Clwyd and MaTRI.
MGN, Mixer, Clwyd each carried out rubber compounding optimising trials to find the best formulation(s) for their applications. Further more details results can be found in 5.3
MaTRI carried out further generic optimisation trials for dissemination purposes.
Rubber Component Manufacturing Trials
Production of rubber components is performed using a great variety of manufacturing techniques and it was considered by the Safe Rubber Consortium to be important to utilise as many methods as possible during the industrial validation trials.
The process starts with mixing, or compounding the elastomer with various ingredients. These include curatives, anti-degredants, processinng aids, fillers and other functional additives. Mixing is carried out either on an open two-roll mill or in an internal mixer..
After mixing the compound must be shaped by methods such as moulding, extrusion or calandering. Once formed into the desired shape the component must be cured, alternatively refered to as crosslinking or vulcanisation.
An industrial trial programme was designed to test SRM102, the newly developed acellerator, as a viable alternative to ETU. Because SRM 102 is currently made in a powder form and used in small quantities of less than two parts per hundred rubber, it is unlikely that it will affect the physical processing behaviour of the compound, for instance mixing and extrusion speeds, dimensional charateristics, heat build up etc. However its cure characteristics could initiate several problems.
These include:
• Premature cure, or scorching, during the mixing process. In its mildest form scorching can take the form of an increase in viscosity and small lumps of cured rubber developing throughout the compound. With the higher temperatures found in an internal mixer, the whole compound can start to cure which can be detected by rapidly rising torque readings
• If scorch is initiated by the heat in an extruder barrel, the result is a rough extrudate
• Scorch during calendering will cause a rough surface finish
• A slow curing extrudate is liable to sag and become misshapen during an autoclave cure
• Porosity in the rubber can occur in a slow curing continuous cured compound
• Mouldings can become distorted if the cure state is incorrect
The Consortium developed a test matrix, which included the types of process, types of product and partners with the capability to perform the trials. It was evident that the Consortium did not have the capability to perform all of the necessary processes. Two companies outside the Consortium offered to provide their assistance with the trials on the basis that the aims of the Project offer great potential to improve safety in the European rubber industry in the future.
Process Application Performing
Extrusion Autoclave cure Round section Nufox Rubber*
Garvey die section Mixer
Continuous cure Round section Nufox Rubber*
Polychloroprene cable sheathing Mixer
Moulding Compression Thin sections MGN
Thick sections MGN
Rubber to metal bonding MGN
Injection moulding Various BD Technical Polymer*
*Non-Consortium Member
Notes:
• As both compression moulding and injection moulding trials were successful, transfer moulding was deemed by the Consortium rubber experts to be un-necessary because it is a hybrid of the two processes.
• Calendering was not trialled on the grounds that it would require a large quantity of compound along with lengthy machine set up times and as such was too expensive. Furthermore the rubber experts in the Consortium agreed that if the extrusion trials were successful, then there would be no reason why calendering would not also be successful.
• Polychloroprene sheathing trials were performed using a specially developed compound to meet industry specifications. This utilised a cross-head extruder.
• Two types of continuous cure were trialled; infra-red and salt-bath.
The manufacturing trials were performed or overseen by four Consortium partners; MaTRI, Mixer, Clwyd and MGN. This report describes the trials by each partner.
Full Analysis of Industrial Scale Validation
As previously mentioned, the Safe Rubber Consortium did not have the capability to perform the full range of trials that was required to validate SRM102 in an industrial environment. MaTRI located two companies outside the Consortium who offered to assist the Project for no fee.
MaTRI provided all partners with guidance and health and safety information. The technical guidance was taken from work performed with SRM102 and SRM104 earlier in the project. The health and safety information was based on work by GSL. The guidance notes are can be found in the appendix od D18.
Injection Moulding
An injection moulding machine consists of a cylindrical injection barrel with a ram or screw inside it. Rubber compound is forced towards a nozzle at its end which is connected to the top half of a closed mould. Within the mould are gates and runners leading to the mould cavity itself. Cure temperatures used in injection moulding can be much higher than those used for compression or transfer moulding enabling faster cure cycles.
Injection moulding trials were performed by BD Technical Polymer, which is a member of the BRPPA. The compound used for the injection moulding trials was the same as Trial Compound 1 used for the continuous cure trials.
A semi-spherical, hollow, parabola-shaped product was chosen for the trials. Its large diameter was approximately 130mm with a height of 60mm. The small end was recessed at approximately 42mm diameter and contained a 10mm hole. The wall thickness was 4mm. The net weight was 175 grams per product. Figure 6 shows the details of the product.
An REP RT9 vertical injection moulding machine was used. The cure temperature was set at 160°C for all trials as being representative of industry practice and also within the operating temperature range of SRM102.
The machine was thoroughly flushed out with the test compound before series of mouldings was produced using a single cavity mould.. The first pieces were moulded for 25 minutes and appeared to be fully cured. State of cure was assessed by general feel and an indentation resistance and recovery test. No moulding defects or abnormalities were identified in the test pieces.
Similar results were obtained after reducing the cycle time to 20 minutes and then to 15 minutes. Moulding for 10 minutes produced samples which were not considered to be fully cured. Further samples were cured using the 15 minute cycle to confirm the earlier results.
The injection moulding trial was considered to be a success. No moulding defects were observed, the compound cured after the mould was fully loaded and there was no indication of premature cure (scorch) occurring. The 15 minute cure cycle could easily be reduced either through optimisation of the accelerator system or by using increased temperatures. The products obtained were of equal quality to products made using traditional ETU accelerated compound.
Continuous cure (non-cable compound)
Two types of continuous cure or vulcanisation (CV) were trialled with compound containing SRM102; salt-bath and infra-red. Both methods use a normal cold feed extruder, the main difference being the method by which heat is provided to the un-cured profile.
Salt bath CV is a commonly used liquid curing method for extrudates. CV is frequently chosen for producing products such as tubing, hoses and weather stripping. Salt baths are relatively short-length curing units because salt has good heat exchange properties and can be used at high temperatures of up to 260°C. Salt does not cause surface oxidation, and is easy to clean off the finished product using water.
Infra-red (IR) CV lines use a row of IR heaters to cure the extrudate and tend to be longer because heat flow into the rubber compound is slower when compared to a salt bath.
In both processes it is essential that the compound starts to cure enough to maintain its shape before distortion due to melting occurs. Another consideration is the formation of porosity which can be caused by the formation of volatiles before cure has taken place. It can also be caused by moisture in the compound turning to steam.
Continuous cure trials were performed by Nufox Rubber on both its infra-red and the salt-bath extrusion lines. Autoclave cure trials were also carried out by Nufox. The trials were supervised by MatRI and Clwyd. Clwyd provided the compound for the trials.
The compound was extruded through a 10mm circular die and various parameters were changed during the course of the trials until optimum conditions were obtained.
Whilst it cannot be considered as continuous cure, samples of extrudate were taken from the extruder line and cured in an autoclave. This process consists of placing the un-cured extrudate on a talc covered, aluminium tray in an atmosphere of pressurised steam. The pressure used was for the trials was 4 bar which equates to 152°C.
Mixer is a manufacturer of rubber and thermoplastic compounds for the insulation, sheathing and filling of flexible cables, rubber cables, power cables and flame-retardant cables. This is a specialist market due to the use of highly specified material with high volumes and low margins, making it different from the general purpose rubber goods industry.
Mixer’s task in the Saferubber project was to validate the use of the new accelerator SRM102, in an industrial manufacturing envoirment and test the substitution of ETU in rubber cables. The overall aim being to reduce the toxicity caused by fumes from ETU during manufacture without losing technical performance.
Mixer’s first task in this part of the SafeRubber project was to validate the use of the new accelerator in industrial manufacturing and to test the substitution of ETU in rubber cables in order to reduce toxicity issues during production without losing technical performance.
SRM102 was previously chosen as the best fit to replace ETU for both rheometric and mechanical properties. This accelerator will be tested as a direct replacement in the compound formulation and as a complete replacement of the whole crosslinking package.
Rheometric, mechanical and ageing properties will be investigated in order to understand if this chemical will be suitable to produce heavy duty, high temperature, oil resistant cables used worldwide for offshore and mining applications.
Five batches were produced and tested in order to check the reliability of the compounding procedure. Just one batch was discharged because it was not fully homogeneous. The remaining four batches were mixed in order to obtain a homogenous amount of compound to be divided and modified with the addition of different crosslinking packages.
During mixing it was quite clear that in the SRM102 sample, lumps of the accelerator were present affecting the dispersion of the chemical.
However the test on the masterbatch was completed in order to show the failure produced by the lumps. Results show lower mechanical and rheometric properties which are caused by poor dispersion of SRM102.
The next set of industrial trials were designed to test the extrusion properties and manufacturing conditions in order to understand the critical features during processing. A standard, currently used, formulation was chosen for the trials: 5GM3 which meets DIN VDE 0207 teil 21.
It was clear that the crosslinking speed of the SRM102 compound is lower (higher t90). For this reason, two new test sets were prepared in order to test the SRM102 concentration effect on the properties, and the best balance of the accelerators in presence of SRM102.
SRM 102 concentration acts as a booster for the vulcanization process increasing the speed and the crosslinking density and mechanical properties are slightly improved when more than 2 phr of SRM102 is used.
In order to understand how SRM102 works together with other accelerators UNIMIB prepared a design of experiments programme. The influence of secondary accelerators was investigated.
TMTM increases compounding and handling safety (increasing ts2) and productivity (reducing t90).
SRM102 and DPG act as standard accelerators reducing ts2, and increasing the crosslinking density.
Sulphur acts similarly to DPG and SRM102 but without reducing t90. Sulphur appears to be the best solution to give increased MH, and for this reason even hardness.
All the experiments performed by Mixer show that SRM102 could be an effective replacement of ETU in polychoroprene based compound.
Crosslinking density, hardness, scorch safety, mechanical properties and more importantly, ageing resistance is very close to what could be obtained with ETU. Crosslinking speed is slower but from the data obtained from UNIMIB’s DoE it is clear that TMTM and DPG are useful additives for tuning the performance of SRM102 for a complete and successful substitution of ETU in heavy duty service cable compounds.
Two compounds were used to compare the effectiveness of SRM102 compared with ETU.
The trial compounds were extruded through an extruder. The extrusion temperatures and pressures were virtually identical for both compounds.
The extruder was fitted with a Garvey die to test the extrusion properties because it was felt that this would highlight any problems such as scorch. It was seen that there is little or no difference between the two compounds.
In conclusion, it can be stated that there is no discernible difference between the processing characteristics of compound containing ETU or SRM102. Both extruded in an identical manner with no problems or issues.
Clwyd Compounders Ltd. was required to carry out industrial evaluation studies on the use of SafeRubber accelerators SRM102 and SRM104 in polychloroprene rubber compounds.
20kg of unfilled polychloroprene masterbatch was mixed and sheeted on a production scale two-roll mill. SRM102 and SRM104 accelerators were added to aliquots of masterbatch on a two-roll laboratory mill at loadings from 0.5phr to 2.5phr blended well and sheeted off.
The rheology of each compound was tested at 185°C for 6 minutes. Cure conditions were based on the cure curves obtained. 2mm thick test sheets and small (6.3mm x 13mm) compression set buttons were moulded of each compound for 15 minutes at 165°C.
Conclusions - cure efficiency initially increases with increased loading of SRM102 but then plateaus at 1.5 phr SRM102. However, cure speed continues to increase at loading of more than 1.5 phr SRM102, increasing scorch risk. Optimal physical, aging and compression set properties were seen at 1.5 phr SRM102. The cure efficiency of SRM104 containing compounds gradually increases with loads up to 2.5 phr SRM104 but so does cure speed. Optimal physical properties, aging and compression set properties are seen at 1.5 phr SRM104.
A study was done to investigate the use SRM102 and SRM104 in commercial polychloroprene compounds by substituting a standard accelerator system for SafeRubber accelerators in a commercial black-filled compound. The resulting compounds were tested against BS 2752 and DTD 5514 standards and results compared against previously obtained results in an ETU containing benchmark compound.
Two compounds 30M3B023 & 30M3B024 were mixed in laboratory internal mixer, then slabbed, re-milled and sheeted on a two-roll laboratory mill. The rheology of each compound was tested at 165°C for 30 minutes. The rheology of each test compound was tested for 30 minutes at moulding temperature (165°C). A 195 x 195 x 2mm test sheet was moulded of each formulation for specification testing against both the BS2752 and DTD5514 specifications.
In conclusion both general purpose and high-quality SRM102 compounds show good basic properties. The high-quality CR compound is within reach of the challenging DTD 5514 specification requirement, which has traditionally required ETU containing compounds to achieve it. Although SRM102 & SRM104-containing general purpose CR compounds show low elongation compared to other accelerator systems, tensile strength and modulus at 100% elongation are higher. Compression set and fluid resistance testing showed comparable results to the standard accelerator systems, suggesting that a good state of cure is achieved in a general-purpose CR compound using 1 phr of SafeRubber accelerators. The short t05 time may lead to shelf life issues, which suggests that shelf life studies on SRM102-accelerated compounds may require consideration.
MGN’s role in Work Package 5 was to validate the use of SRM102 in moulding compounds an industrial environment, specifically compression moulding of various sized parts and rubber to metal bonding. MGN does not possess any injection or transfer moulds. MGN selected some pieces currently manufactured by the company with chloroprene rubber in production moulds, in order to test the whole range of applications described earlier.
The work to be done is to manufacture all these parts, substituting the current ETU containing compound used by MGN for a new compound, having the same formula (or as similar as possible), but substituting ETU with the new accelerator developed in the SafeRubber Project, SRM102.
The industrial validation of the suitability of the new accelerator for its use in the vulcanization of CR consists on the comparison of several significant parameters, with those obtained with the current CR used by MGN:
MGN uses basically two different CR blends, so before vulcanizing the chosen references, a previous step had to be done, to substitute the ETU by SRM102 in those two blends. During this step, various combinations in the formulations where tried to introduce SRM102 in MGN’s CR blends, having as reference original MGN CR blends to compare the results obtained.
The study was performed in two steps:
• Step one: a formulation was chosen and some trials on real moulds were made.
• Step two: Some more versions of the formula were made, trying to improve the vulcanization performance and one was chosen
General conclusions from the vulcanization curve:
• Suitable viscosity in all of them
• Enough scorch time in all of them
• Slower vulcanization rate when using SRM102 at tested amounts (up to 1.8 phr; if more used, insufficient scorch time would appear)
• Similar level of vulcanization than the original one can be achieved depending on the SRM102 phr
• The use of TMTM instead of TMTD doesn’t improve significantly the vulcanization
MGN has optimized its two CR formulas to be used substituting the ETU by the new accelerator SRM102, and has tried them in the moulds proposed in order to check all the technologies and applications described in Table 27.
The new CR formulas with SRM102 have been found to work properly in all the tested technologies and applications.
In a first set of trials, it seemed that the use of the new rubbers could arise in a loss of productivity, as the vulcanization curves suggested that a little extra time should be used for vulcanization. However, a second set of trials was successfully conducted to check and prove that the new rubbers could be used in the tested moulds with the standard vulcanization time, so the use of these doesn’t imply any productivity loss.
In addition, some characteristics of the new rubbers have been found to improve the ones of the standard formulas:
• New formulas have higher tensile strength than the standard ones, so a characteristic as important as the tension at break in a piece like the guiding part tested, results to be higher if the piece is manufactured with the new CR.
• Along the experience with the vulcanization trials, the ability to flow and fill easily the mould has been found to be much better in the CR rubbers made with SRM102 than in the standard ones. This could arise in a reduction of the mass of rubber needed to vulcanize a piece
As a general conclusion, it can be said that the industrial validation of the new CR rubbers made with SRM102 has been successfully achieved, for all the technologies and applications tested by MGN.
Chloroprene compounds containing SRM 102 have been tested in a wide array of industrial processing techniques in five factories in three different countries. In every case SRM102 has equalled or exceeded compound containing ETU. It has proved itself to be a viable replacement for ETU as an accelerator.
Commercialisation
RBL is unable to estimate the definitive price of SRM102 until it produces the product on an industrial scale (it has only produced 44 kg so far). At the moment is not possible to compare the current cost of MGN’s standard formulas with the one of the new formulas, although the variation should be very small, due to the small proportion of ETU and SRM102 used in the formulas.
Potential Impact:
The SafeRubber project has provided the SME members with an IP protectable method (Trade Secret) to manufacture chloroprene rubber, which will enable them to effectively compete with large enterprises and Asian importers. The European market for chloroprene rubber is over €210 million per annum10 and the Global market ~€650 million10 per annum. Assuming a modest market penetration of 10% by 2017 (five year post project), this equates to increased sales (displace imports) of €45.3 million cumulatively for the SME members. This extra €41.5 M turnover will help safeguard or create 453 jobs12 within Europe.
In addition to this, it has been conservatively estimated that a market penetration of 3% by 2017 outside the EU (exports) equates to sales of €36.4 million creating or safeguarding an extra 364 jobs within Europe. This project also supported the SMEs who supply the chemicals and processes to distribute the product developed to the rubber industry.
A replacement that is safer than ETU has been developed i.e. SRM 102 which has proven to be a more environmentally friendly process, which in turn leads to inherently fewer hazards and less arduous health and safety requirements. The accelerator has also been successful in processes i.e. epichlorohydrin rubber (ECO).
SafeRubber has produced a more environmentally friendly process, which in turn will lead to inherently fewer hazards and less arduous health and safety requirements.
During the moulding trials it was proved that compound made with the new accelerator also had better flow properties, this should result in less scrap being produced, hence less landfill. This has been industrially validated by all the end user beneficiaries with comments that this has equalled if not exceeded that curing properties of ETU.
The process has developed a platform technology that has been proven that it can be applied to other processes for example, epichlorohydrin rubber (ECO).
Additional trials have also shown that it is possible to replace/reduce the ZnO which has a significant environmental benefit by using a multi-functional additive or TiO2 thus the environmental impact can also be reduced through significantly lowering the quantities of MgO and ZnO required in the process – conserving raw materials and reducing the quantities of hazardous effluent.
The project has been successful and now means that there is an accelerator which is suitable for the production of poly chloroprene and is now not gender restrictive.
Considerable interest has been shown in the new molecule which should benefit both the consortium and the European rubber industry
List of Websites:
www.saferubber.eu
Coordinator - Girolamo Dagostino - g.dagostino@assocomaplast.org
Technical Manager - David Cartlidge - david.cartlidge@pera.com
Dissemination Manager - Geert Scheys - gscheys@essenscia.be
A review of published work was undertaken to compare the already-suggested mechanisms and to further guide subsequent experimental work. In order to design better curing agents for chloroprene rubber, studies of the curing mechanism using advanced analytical techniques and computational methods continued. It has been found that, as expected, the mechanism of curing changes depending on the chemical used as the crosslinking agent. Our results lead us to the conclusion therefore that the mode of action of ETU cannot simply be described as one mechanism, but rather is a combination of multiple crosslinking reactions which change depending on the reaction conditions employed.
From the desk-based review of toxicity the key message is that ETU, the present used accelerator, has been classified as Reproductive toxicant category 1B with Hazard statement H360D – may damage the unborn child. Being classified as CMR, ETU is a substance that could be identified under REACH as a Substance of Very High Concern (SVHC) and therefor a potential candidate to be included in the Candidate List and later in Annex XIV of REACH (Authorisation).
The REACH Regulation mentions (Q)SARs as a mean to provide data without doing animal testing. ECHA (European Chemical Agency) presented a paper in 2009 informing that results of QSAR may be used instead of testing when certain criteria are met. As part of Annex VII testing for REACH registration (volume 1-10 t/y) and also Article 26 notification to ECHA is important. This work is currently on-going.
A life cycle assessment was carried out using the 4 steps: Goal and scope definition, inventory Analysis, Impact Assessment and Interpretation. Also tests to reduce/replace Zinc Oxide in the compound were carried out as well as leaching trials. Some findings from these tests relevant for environmental impact are
• the ability to flow and fill easily the mould has been found to be much better in the CR rubbers made with SRM102 than in the standard ones. This could arise in a reduction of the mass of rubber needed to vulcanize a piece which in turn will reduce the scrap.
• The cure cycle could easily be reduced through optimisation of the accelerator system
• It is possible to reduce the level of ZnO, by using either TiO2 or MFA’s especially when SRM102 is used as the accelerator.
• After 12 weeks leaching samples look similar. However, algae growth on samples produced by use of SRM102 suggests they’re not toxic!
Industrial validation trials were carried with chloroprene compounds containing SRM 102 and were tested in a wide array of industrial processing techniques in five factories in three different countries. In every case SRM102 has performed as well as, or better than compound containing ETU. It has proved itself to be a viable replacement for ETU as an accelerator. The rubber manufacturing trials have proven that the SRM102 is also a commercially feasible replacement for ETU. All industrial validation trials carried out during the project indicate that the proportion of SRM102 used in the formulas for all the end users will be very similar to ETU i.e. around 1%, therefore the current price estimates of the accelerator SRM102 are not anticipated to increase the manufacturing cost of a final rubber compound more than 105% of the equivalent compound made using ETU.
Dissemination has been active throughout the project and high visibility achieved for project partners and project results resulting in: More than 41 press quotes, 9 articles in rubber technology & science magazines (1 peer reviewed) + 1 master thesis dissertation, 4,890 unique visitors of the project website until May 2013. Dissemination of information by consortium members at 61 EU and global rubber industry events (conferences & trade fairs): 832 participants at conferences and a potential of 700,000 trade fair visitors. Presentation of project results at 5 Rubber Technology and 2 Chemometrics seminars in UK, Belgium, France and Italy. Training achieved of all project partners on the dissemination of the project results by the interactive development of a Dissemination Presentation Tool
Project Context and Objectives:
The primary objective of the SafeRubber project is to develop a commercially viable replacement accelerator molecule that can effectively replace thiourea based accelerators and metal oxides in the vulcanisation of rubbers such as chloroprene rubber. This is to be accomplished through the development of a new multifunctional additive and will be developed for use in existing processing equipment with subsequent development work taking into account processing conditions to provide an accurate grading of the resultant product, in terms of physical and chemical properties.
In order to achieve this overall aim, the consortium need to overcome a number of technical hurdles, which have defined a list of key objectives listed below:
1.To develop a replacement accelerator molecule (based upon new molecule or modification of existing molecule) that can produce polychloroprene rubber that performs to within ±10% of the parameters obtained with existing chemistries (thiourea based accelerators) such that the final product must:
a.Cure at a temperature and time similar to that obtained with the ETU/CR process (comparable Mooney Scorch curve at 130°C in accordance with ASTM D1646 or equivalent standard)
b.Enable materials to be produced with a tensile strength in the region 15-25 MPa, compression set of 15% after 24hrs at 70°C, modulus at 100% of 5.6 MPa and hardness of 74 IRHD (ISO 48:2007 or equivalent standard)
c.Have an operating temperature range of -45 to 120°C
d.Have excellent heat, oil and chemical resistance (in accordance with ISO 2475 or equivalent standard)
e.Not scorch or discolour during rubber processing (in accordance with ASTM D2084 or equivalent standard)
f.Have good UV, ozone and weathering resistance (in accordance with ASTM D0750-06/D1148-07 or equivalent standard)
g.Have an odour during processing and use that is no worse than ETU, as assessed by RBL’s Odour Panel.
2.To develop a safer alternative accelerator to ETU that complies with REACH and European Health and Safety requirements and is commercially viable at less than 105% of the equivalent product made using ETU.
3.To develop a scalable process for the large-scale production (~200 tonnes p.a.) of the replacement accelerator molecule such that the manufacturing cost of a final rubber compound is less than 105% of the equivalent compound made using ETU.
4.To demonstrate a scalable process by producing a large quantity (1-10 tonnes) of replacement accelerator molecule for use with Industrial Scale Validation trials.
5.To conduct industrial scale validation trials to demonstrate the performance of the replacement accelerator curing system in comparison to existing products and formulations on a industrial scale.
In addition to these project objectives, there are two further valuable contributions which the project will achieve:
a)Enhance the understanding of the fundamental mechanism of action of the accelerator system, based upon Ethylene Thiourea (ETU) and MgO/ZnO, which will allow further advances in the general understanding of the curing of polychloroprene and associated rubbers.
b)Enhance the understanding of using QSAR methodology to provide information on the feasibility of potential molecules through understanding how the molecular structure will affect toxicity while maintaining performance properties, especially performance properties especially as the current system is not based on an API (active pharmaceutical ingredient) for which QSAR is traditionally used.
Project Results:
Development of New Active Accelerator Molecule
Mechanistic Study
In order to design a suitable ETU replacement in the Saferubber project, the precise mode of action of ETU must be understood. To begin to understand this, a review of published work was undertaken to compare the already-suggested mechanisms and to further guide subsequent experimental work.
The curing of chloroprene monomers was next investigated. A significant delay was experienced at this point as the chloroprene monomer was not commercially available and had to be synthesised in the laboratory. The monomer was found to be quite unstable, undergoing gradual self-polymerisation over a number of days. In addition, cross-linking experiments with ETU could not be completed as the low boiling point of the monomer meant polymerisation and/or decomposition occurred before reaction with ETU was observed.
In parallel with Task 1.3 initial investigations into the curing of low molecular weight oligomers of chloroprene were also carried out to gain an understanding of the reactivity of the material. The initial results from this experimental work using FTIR and NMR analyses as suggested in the Description of Work did not yield a definitive answer to the mode of action of ETU as a wide spectrum of chloroprene mixtures was not prepared. However, the experiments have provided enough information to permit appropriate experimental design in the subsequent tasks and to suggest alternative analytical techniques for Tasks 1.2 and 1.3.
In order to design better curing agents for chloroprene rubber, we have continued to study the curing mechanism using advanced analytical techniques and computational methods. Using both chloroprene oligomers and gumstocks, Fourier Transfor infra-red (FTIR) spectroscopy and gas chromatography-mass spectrometry (GC-MS) have been employed to probe structural changes in the compounds to try to identify what is happening during crosslinking.
Kinetic studies show that zinc oxide (ZnO) rearranges the 1,2-isomer before crosslinking takes place. FTIR spectroscopy was used to monitor the disappearance of the band at 925 cm-1 (attributed to the C=C group in the 1,2-isomer) during crosslinking without additives and with ethylene thiourea (ETU), ZnO, and both ETU and ZnO together. When polychloroprene gumstock was heated to 160°C (without additives) there was a zero-order depletion of the 1,2-isomer during the first 25 minutes, reduced by over 90%. The remaining isomer was removed more slowly in the subsequent 10 minutes. This shows that the rearrangement of the 1,2-isomer can occur slowly on heating without any additives. With ETU as the only additive, the 1,2-isomer reduced much quicker. Approximately 60% was removed within the first 60 seconds. The remainder of the 1,2-isomer steadily depleted over the final 25 minutes. This confirms that ETU rearranges the 1,2-isomer quicker than when no additives are used at all. When ZnO is used as the only additive, the 1,2-isomer rearranges faster than under any of the other tested conditions, completely disappearing within a minute. With both ZnO and ETU as additives, the rearrangement is fast. There is a 90% reduction of the 1,2-isomer within one minute and the remainder disappears within 10 minutes. This is slower than with ZnO alone and suggests that there is a strong interaction (or complexation) between ZnO and ETU, which competes with the reaction of ZnO with the 1,2-isomer.
Systematic studies on the crosslinking of polychloroprene confirm that there are a number of competing mechanisms occurring at varying rates, rather than one single, straightforward reaction. The various reactions that have been identified include; (i) polychloroprene cross-linking with ETU through both its amine groups, as postulated by Kovacic (1955), (ii) cross-linking with ETU through one amine group and one sulphur, in the presence of ZnO and (iii) cross-linking of sulphur from ETU (in the presence of ZnO), which sees the formation of ethylene urea (EU) as a by-product. A possible route for the latter mechanism was published by Pariser (1960). Our studies have revealed evidence for each of these, or similar, competing mechanisms. When ETU and ZnO are both used, EU is evolved as a by-product, as confirmed by FT-IR and GC-MS, which is strong evidence for the Pariser mechanism. However, the published mechanism indicates that ETU reacts first with polychloroprene, before ZnO. However, we have strong evidence against this, as our kinetic study shows that ZnO reacts significantly faster with polychloroprene. Our computational analysis work also disagrees with the published mechanism and supports our own proposed mechanism.
Based on our current understanding, diamines such as 1,4 diaminobutane in combination with tetramethyl thiuram disulphide (TMTD) (as a less toxic model for ETU) were trialled as potential candidate accelerators. As predicted, this new accelerator combination was successful in crosslinking polychloroprene. The presence of the diamine appears to activate the allylic position of the 1,2-isomer on the polychloroprene chain in a similar manner to ZnO, albeit via a different mechanism, through which TMTD is then able to cross-link. It is possible that HCl released becomes incorporated as the diamine HCl salt. It should be noted that the diamines are also able to crosslink polychloroprene alone albeit poorly, as observed from results obtained through rheology and mechanical testing. The above observations give further confirmation of our understanding of the mechanisms occurring and allowed us to design new, safer accelerator salts containing anionic sulphur centres and cationic diamines that have been shown to be highly active for the curing of chloroprene.
Based on our current understanding, diamines such as 1,4 diaminobutane in combination with tetramethyl thiuram disulfide (TMTD) (as a less toxic model for ETU) were trialled as potential candidate accelerators. As predicted, this new accelerator combination was successful in crosslinking polychloroprene. The above observations give further confirmation of our understanding of the mechanisms occurring and allowed us to design new, safer accelerator salts containing anionic sulphur centres and cationic diamines that have been shown to be highly active for the curing of chloroprene.
During the final period of the project, work has continued in WP1 to understand the mechanism of curing of chloroprene rubber. This has built on the knowledge already gained on the mode of action of ETU from the previous periods of the project. The in-depth analysis of FTIR spectroscopy traces obtained from cured and non-cured oligomers and gumstocks has continued in order to identify the salient peaks and what they are caused by. Although the FTIR data had given us a deep understanding of the curing mechanism, the absolute identity of the peaks in the 1540 cm-1 region remained unknown. These are critical to understand, as they are attributed to the actual chemical crosslink in the polymer. Using chemical mimics for ETU, much work has been undertaken to observe the effect of the curing conditions on the peak sizes and positions in FTIR spectra. Tensile and other rheological data has then been obtained for the compounds cured with the chemical mimics to observe the effects caused by the replacement of ETU. Elemental analysis has been used as a probe for the chemical composition of cured products.
It has been found that, as expected, the mechanism of curing changes depending on the chemical used as the crosslinking agent. However, it has also been noted that proposed mechanisms of curing do not act independently and the FTIR data and tensile testing results suggest multiple cross-linking reactions take place, depending on the curing agent used; when zinc oxide and thioureas are used, bisalkylation dominates cross-linking; when thioureas or similar compounds are used, sulphur cross-linking dominates; when bases are used, cationic crosslinking dominates. Our results lead us to the conclusion therefore that the mode of action of ETU cannot simply be described as one mechanism, but rather is a combination of multiple crosslinking reactions which change depending on the reaction conditions employed.
To finally identify the peaks observed by FTIR spectroscopy in the 1540 cm-1 region, it has been proposed that Raman spectroscopy should be used in place of FTIR spectroscopy. It is believed that Raman spectroscopy may be able to distinguish between carbon-nitrogen and carbon-sulphur bonds, which would allow us to identify if bisalkylation or sulphur-crosslinking dominates in a particular reaction. However, Raman spectroscopy is a specialised technique and not widely available. A collaboration is being set up with a supplier of the technique and this work will continue outside of the project.
Confirmation of mechanism using gumstock resin
Gumstock resins were studied in this task in order to compare analytical results to the data obtained from the low molecular weight oligomers and the literature data. A set of gumstock samples was prepared containing various combinations of ETU, zinc oxide and magnesium oxide and cured under standard conditions. These were: gumstock and ETU; gumstock and ETU, MgO and ZnO; gumstock, MgO and ZnO; gumstock, ZnCl2 and ETU; gumstock, ZnO and EU (2-imidazolidone, ethyleneurea); gumstock and EU; gumstock and ZnCl2. Structural variations in the different uncured and cured rubbers were then investigated using solid-state FTIR spectrophotometry.
However, these data were not conclusive as only very minor variations in the FTIR traces were observed for each different sample; this was not unexpected as the structural changes are subtle and the curing agent is only present in very low concentrations. A wide range of other analytical techniques was therefore undertaken on the curing and cured gumstocks in order to support or reject suggested curing mechanisms and supply additional data.
In addition, computational analysis has been used to help validate and/or contradict the literature curing mechanisms. This analysis uses computational chemistry software on a supercomputer to predict likely transition states and intermediates during chemical reactions. The overall conclusion from this preliminary analysis was that the suggested mechanisms for the curing of chloroprene in the literature were possible but contained a number of unlikely transition states and intermediates. The information obtained from these computational analyses was combined with the experimental data to help explain observations and lend further weight to predicted curing mechanisms.
Molecular design of replacement accelerator and lab scale trials
QSAR modelling of replacement accelerator molecules
A thorough literature review was carried out to understand how and the extent to which QSAR has been used to estimate accelerator properties. It concluded that so far the use of QSAR to this end has not been exploited. A comprehensive bibliographic collection, which currently comprises 105 scientific papers related to different aspects of the world of rubbers, was generated. We retrieved from literature information about the characteristics of the vulcanization process itself (such as scorch time, optimum cure time, cure rate index, cure temperature, maximum and minimum torque) and the mechanical properties of the vulcanized (tear and tensile strength, elongation at break, hardness, crosslink density, modulus %, swelling index).
Documentation on available experimental data for existing accelerators: mechanical and rheological properties of rubber samples collected from the scientific literature were organised into a database. It includes 950 rubber samples (with different types of accelerator, formulation, rubber and temperature of vulcanisation), 11 mechanical and rheological properties measured on each sample, 5 qualitative variables defining type of accelerator, formulation, rubber, bibliographic source and vulcanisation temperature.
Statistical analysis of the mechanical and rheological properties of different compounds: the mechanical and rheological properties of different compounds were analysed by means of multivariate statistical tools (Principal Component Analysis) in order to understand the role played by different factors (type of accelerator, type of formulation, type of filler, type of rubber) in defining the rubber properties. A pattern mainly related to the bibliographic source was discovered. This suggested that a) experimental conditions/procedures really affect the measure of mechanical and rheological properties of compounds; b) the inter-laboratory variance seems to be the major source of discrimination of compounds. For this reason, a set of accelerators to be tested for their performance in vulcanization has been selected by RTD partners to obtain reliable data for modelling.
QSAR studies for prediction of curing properties: QSAR models attempt to relate structural features of accelerators (encoded by molecular descriptors) to their curing properties in order to predict properties of untested accelerators and select the best accelerator candidates. In order to set up the QSAR analysis, first the properties to be modelled were defined; these are optimum cure time, scorch time and cure efficiency. Then, a large set of 1650 molecular descriptors expected to be related to the curing properties was selected. Algorithms for the calculation of these molecular descriptors from the chemical structure files were generated. The most suitable multivariate statistical strategies to select the optimal subset of molecular descriptors for curing property calculation were evaluated. MATLAB codes to implement the above mentioned strategies and those for the calculation of mathematical models were generated. Finally, the molecular structures of the existing accelerators under analysis were designed and optimized by specific chemical drawing applications.
QSAR studies for prediction of toxicological end-points: existing QSAR models that relate chemical structures of substances to toxicological properties were evaluated and their efficiency was compared. The most effective models will be used to predict toxicological properties of accelerator candidates. Activities were initially focussed on a thorough review of documents on toxicological end-points relevant to REACH, along with QSAR scientific papers, shared by NORNER. Then, we carried out a screening of existing QSAR applications (i.e. CAESAR, TOXTREE, LAZAR, EPISUITE, QSAR TOOLBOX, SPARC, PBT PROFILER, VCCLAB, ONCOLOGIC), we evaluated toxicological end-points provided by each, the reliability of the predicted toxicological values, as well as their applicability domain. We mainly focussed on persistence, bioaccumulation, toxicity, toxicological, ecotoxicological and physicochemical properties, as suggested by NORNER and recommended in REACH annexes XIII and VII, respectively.
Existing QSAR models that relate chemical structures of substances to toxicological properties were evaluated and their efficiencies were compared. The most effective models were selected to predict toxicological properties of accelerator candidates. Activities were initially focussed on a thorough review of documents on toxicological properties relevant to REACH. The following properties were taken into considerations, since they were suggested by NORNER or recommended in REACH annexes XIII and VII, or commented in the SafeRubber project: bioaccumulation, biodegradability, carcinogenicity, skin sensitization, mutagenicity, aquatic acute toxicity, acute oral toxicity, developmental toxicity. A screening of existing QSAR tools (CAESAR, ToxTree, Lazar, Episuite, OECD QSAR Toolbox, SPARC, PBT profiler, VCCLAB, Oncologic, T.E.S.T) was carried out. Toxicological end-points provided by each tool, reliability of predicted toxicological values, as well as definition of applicability domain in each model were evaluated. CAESAR, ToxTree, OECD QSAR Toolbox, Lazar, Episuite and T.E.S.T. were selected as the most reliable QSAR tools and used for subsequent toxicological profiling.
ToxTree toxicological models are mainly based on a compilation of structural alerts and rules. These rules were analysed in order to enhance the selection of safe accelerators. Results of this analysis were shared with SafeRubber partners in November 2011. While analysing the ToxTree rules, some divergences between actual predictions and expected predictions were found. UNIMIB had a continuous contact with the ToxTree software webhouse in order to clarify these issues. It was found that some predictions provided by the Structural Alerts for the in-vivo micronucleus assay in rodents model according to rule “QSA34; H-acceptor-path3-H-acceptor” were erroneous; related predictions were thus manually changed and the bug was officially reported in the bug list of ToxTree. Moreover, the model for skin sensitization potential was modified in release v2.5.1 following discussion with ToxTree development team and detection of a bug on rule “Schiff base formation”. This rule includes several potentially sensitizing fragments, one of which being defined by the SMARTS number 12 [CH2][NH2]. In the developer’s intention, this SMARTS should detect only primary amines, but also secondary protonated (tetravalent) amines were erroneously detected.
Since toxicological screening dealt with several salts and some of the considered QSAR models could not give predictions for disconnected structures, corrective actions were undertaken. Initially, commercial QSAR softwares were evaluated (ACD/LABS, Simulations Plus, Accelrys), but none of them could support disconnected structures. Queries were submitted to the databases of OECD QSAR Toolbox in order to find disconnected structures for read-across implementation. Only the most similar compounds to the target salts were retained and this resulted in a total of 24 salts. This small set of disconnected structures was considered not appropriate to implement reliable read-across modelling. Finally, no publications dealing with QSAR models on disconnected structures were found in the scientific literature. As corrective action, it was decided to proceed with toxicological profiling of each ion present in disconnected structures, taken as separate entity.
QSAR models attempt to relate structural features of accelerators (encoded by molecular descriptors) to their curing properties in order to predict properties of untested accelerators and therefore provide information for the selection of the best accelerator candidates. In order to set up the QSAR analysis, first the properties to be modelled were defined; these were TS2 (scorch time), T90 (optimum cure time), minimum torque, maximum torque, the binary variable cured/not cured, tensile strength, modulus 100%, modulus 300%, elongation at break, hardness, as defined in the SafeRubber Annex I and in the project meetings of project Period 1. Then, a large set of 1650 molecular descriptors expected to be related to the curing properties was calculated. Algorithms for calculation of these molecular descriptors from chemical structure files were generated. The most suitable multivariate statistical strategies to select the optimal subset of molecular descriptors for curing property calculation were evaluated. MATLAB codes to implement the above mentioned strategies and those for the calculation of mathematical models were generated. Finally, the molecular structures of the candidate accelerators under analysis were designed and optimized by specific chemical drawing applications.
In order to get reliable QSAR models for curing properties of accelerators, experimental data for a set of commercial accelerators were provided by MaTRI. Ad-hoc QSAR models were calibrated on these experimental data. Due to linear correlation, the property Modulus 300% having large correlation with Modulus 100% was disregarded since equivalent QSAR models would have been found. For the same reason, cure efficiency, which was correlated to Maximum Torque, was excluded from the QSAR modelling.
Selection of replacement accelerator molecules
In order to produce the list of accelerator molecules to take forward to Task (Synthesis of accelerator molecules), it was necessary to define the chemical space that included potential accelerators and all the information needed to find a replacement accelerator molecule with similar performance to thiourea based accelerators, while avoiding their negative toxicological effects towards human health and environment. Initially, potential accelerator candidates were selected and included in an extended library on the basis of the following procedures: a) library design on chemical scaffolds suggested by the SafeRubber consortium and b) screening of “safe” molecules in toxicological databases through structural similarity analysis to existing accelerators. The mode of action of the accelerator system was not exactly defined in WP1 (Development of New Active Accelerator Molecule). Nevertheless, RBL from their experience suggested 12 new molecules and these were also chosen for analysis into their suitability as alternatives for ETU. Multifunctional additive materials, based upon bis-amino compounds containing a thione group (C=S bond) and an additional S–S linkage in the molecule, were selected for further study as potential new accelerators for the curing of chloroprene rubber. Upon heating, the multifunctional additive breaks down to yield a bifunctional amine and a thiuram disulfide species whereby it is believed the bifunctional amine will act as an accelerator for the vulcanisation process and the thiuram disulphide species will act as a cross linking agent.
Evaluation and definition of the criteria to select the final accelerator candidates: The following factors were defined to be relevant to the selection of replacement accelerators: safety/toxicity, curing and mechanical performance, potential economic cost of synthesis, potential feasibility of synthetic route. For each of these factors, a screening aimed to find the most specific indicators for accelerator selection was carried out. Indeed, for each toxicological property, a QSAR-based score was derived. Experimental values, if available, were considered instead of the QSAR predictions. Toxicological scores, mechanical, rheological, physicochemical properties and the indicators of economic cost and feasibility of synthetic route were evaluated for each molecule in analysis and finally combined to define a global accelerator score by means of Multicriteria Decision Making strategies. This global accelerator score accounted for all the relevant information of each molecule and allowed the molecules to be ranked from the most to the least preferable ones.
Prediction of toxicological and curing properties: All of the aforementioned indicators were evaluated for each molecule included in the extended library of potential accelerator candidates. This work was undertaken computationally by applying QSAR tools or through literature review in order to reduce time and costs of analysis and avoid animal testing. Molecules in the extended library were finally ranked on the basis of the selection criterion derived from the predicted indicators by means of Multicriteria Decision Making strategies. Only the first potential accelerator molecules were retained for further evaluation. Toxicological profiles based on QSAR predictions were also evaluated for all of the molecules proposed on the basis of experts’ knowledge about vulcanisation process in order to ensure toxicological safety of the final selected accelerators.
The final list of accelerators to take forward was selected considering both the molecules selected from the extended library and the molecules proposed by experts within the SafeRubber consortium. This selection comprised 15 molecules and was carried out by a panel of experts within the SafeRubber consortium. These molecules were numbered SRM100 (SafeRubber Molecule 100) upwards for identification purposes. The final 15 molecules were submitted to a deeper toxicological evaluation based on both QSAR models and a thorough search for experimental data by means of the available toxicological databases. More QSAR models were considered at this stage, since in the meanwhile several QSAR toolboxes released new models. Finally, the toxicological profile of ETU was compared with those of the 15 SRM accelerators. All 15 selected accelerators can be considered potentially safer than ETU, this being classed as carcinogen and mutagen on the basis of experimental results and classed toxic to reproduction in the CLP regulation
The majority of SRM molecules (14 out of 15 available molecules) are disconnected structures (salts). Only Lazar and ToxTree models can deal with predictions for salts. The toxicological properties which could not be predicted by means of QSAR models for SRM salts due to the lack of suitable QSAR models. Moreover, details, amendments and comments on the toxicological profiles of SRM accelerators were derived and shared with the project Partners. In particular, four molecules were predicted by the Structural Alerts for the in-vivo micronucleus assay in rodents model of ToxTree as being mutagenic according to rule “QSA34; H-acceptor-path3-H-acceptor”. The following considerations were drawn: a) this rule has a True Positive Rate (TPR, percentage of correct positive classification) equal to 34%. This is the lowest TPR in the decision rule of the ToxTree model and it derives that under this rule the considered QSAR model tends to classify as mutagenic many non-mutagenic molecules; b) the two considered Lazar models on mutagenicity gave for the same molecule non-mutagenic result; for these reasons, the four SRM molecules were considered as potentially non-mutagenic molecules. Regarding the ToxTree carcinogenicity model, one of the rules that identifies thiuram as carcinogenic molecule is a nitrogen atom (connectivity 3) bonded to a carbon atom, double bonded to S (connectivity 1) and to any other group but O or S bonded to C sp3, OH, SH, S- and O-. It derives that SRM salts are not identified as carcinogens, but the corresponding covalent molecule would be identified as carcinogen. This information was shared with the project Partners and further detailed in Deliverable 5QSAR methods were applied in order to analyse how molecular structure of Chloroprene rubber accelerators relates to their rheological and mechanical properties. QSAR models were developed in order to disclose which structural features mainly affect the mechanism of vulcanization. Regression mathematical models were developed for two rheological properties (Scorch time and Optimum cure time), and three mechanical properties (Modulus 100%, Hardness and Elongation at break). QSAR models were calibrated using experimental values of fourteen accelerators belonging to diverse chemical classes. A structural interpretation of each QSAR model was given, drawing hypotheses on the correlations between specific structural features and the analysed rheological and mechanical properties as well as defining possible chemical patterns connected to the mechanism of vulcanization.
Task 2.3 Synthesis of accelerator molecules
Nine of the twelve proposed molecules were synthesised as possible accelerators. A decision to leave two of the molecules out was taken as they were in the Class C category and would be included later if no suitable accelerators were identified in the Class A and class B molecules.
The synthesis of each molecule was carried out on a small laboratory scale up to 20 grammes and samples sent for curing trials.
Task 2.4 Lab-scale rubber compounding trials
The first phase of work involved testing many existing types of accelerator in CR compounds in order to build up a thorough knowledge of their effects on cure and to enable input into the QSAR modelling. Some of these have been used in CR previously and some have not. Both single accelerators and combinations of accelerators were tested. Additionally other chemicals, such as sulphur for instance, were used in some formulations to assess their effect on cure.
The second phase involved obtaining benchmark data for ETU so that comparisons can be made between it and the new SafeRubber molecule later in the Project.
Phase three consisted of testing all of the newly synthesised replacement molecules to assess their suitability for ETU replacement. At the conclusion of phase three, two molecules were identified as potential candidates suitable for up-scaling.
Laboratory scale industrial rubber processing and testing equipment was used throughout the trials.
All compounds were mixed using a 300 mm two-roll mill with a friction ratio of 1:1.5. The rolls were temperature controlled using the same setting of 20° C, front and rear, for all mixes. The CR was introduced to the machine on a tight nip and milled until a smooth band of rubber was formed on the front roll. The nip was then opened such that a full width rolling band was created in the nip to enable fast and efficient mixing. Because the compound temperature was low during mixing, all ingredients were added at the same time. Once the ingredients were dispersed the compound was cut off the mill and placed in a water bath to cool to ambient temperature. The compound was not left in the water bath once it was cool as CR compounds are known to absorb water.
To accommodate different size batches, the ears (side plates or guides) can be adjusted to alter the effective width of the rolls. The minimum batch size that can be mixed is around 100g, with the maximum batch size approaching 500g. During the SafeRubber Project, the largest batch that was mixed was approximately 275g.
If a compound proved to be capable of cure, test sheets and hardness slabs were then moulded using a 300mm x 300mm, electrically heated, hydraulic, up-stroking press. Compression moulds were used, one to produce hardness test buttons 30mm diameter x 20mm high, and the other to produce test sheets 220mm x 220mm x 2mm. The sheets were used to provide dumbbell shapes for tensile and other testing.
After mixing, each compound was tested on a Monsanto 100S rheometer to obtain its cure characteristics. The 100S is an oscillating disc-type rheometer. The rheometer plots a graph of torque verses time for any given cure temperature. The full extent of cure and beyond can be recorded. For example reversion, the point at which the cured compound breaks down due to prolonged heating can be observed and measured. Labline software was used to record and interpret the results.
Unless it was known from previous experience that a higher temperature was required for a particular accelerator, all test compounds were rheometer cured at 160° C to represent a typical industrial cure temperature. The remainder were cured at 175° C. The angle of arc was set at 3° with a frequency of 1.7 Hz.
A Testometric FS300CT tensile testing machine was used to carry out a range of tensile related tests: tensile strength, modulus at 100 and 300%, and elongation at break.
Hardness was measured using a Wallace bench-mounted, vibrating, hardness tester or a Wallace handheld hardness meter. Both were calibrated and gave identical results.
To obtain the maximum value from the trials, it was essential that variables between batches were minimised wherever possible. To provide a consistent baseline material for the laboratory trials, a 20 kg batch of masterbatch was mixed in one lot by Clwyd Compounders, using its production equipment. Although each trial batch still had to have the different accelerators added to it individually, this method reduced the risk of deviation due to inconsistent weighing of the core ingredients and variations in mixing.
After addition of the trial accelerators, the compounds were rheometer tested to establish their cure characteristics. Those which cured were then moulded and their physical properties assessed; hardness, tensile strength, modulus at 100% and 300% (if applicable), and elongation at break.
An indication of cure efficiency was obtained by deducting the figure for the minimum rheometer torque from the maximum torque. A low figure showing little cure and a high figure a good state of cure. From examining the results of the trials and comparing them to well-known parameters such as those of ETU, it can be stated that a cure efficiency of over 78 shows a good state of cure.
The first phase of compounding trials consisted of conducting a series of trials to establish the effects of a range of accelerators on the cure characteristics of a gumstock CR compound. Some accelerators were used in the masterbatch alone and some with the addition of a second accelerator or other curative. Along with others, all of the main rubber accelerator groups were represented whether or not they are generally used in CR compounds.
The information gained was used during the initial design of the new SafeRubber molecules. The effects of different molecular structures on the curing of CR could be assessed and taken into account when drawing up a shortlist of potential candidates.
If a compound showed little or no cure, physical testing (tensile, modulus, elongation at break, hardness) was not performed. In these cases, further trials were completed with higher concentrations of accelerator to establish whether or not they made a difference.
To enable comparisons to be made between the new SafeRubber accelerator and ETU, a series of six compounds was mixed for use as a benchmark. 1phr of ETU was used in the benchmark compounds because this is representative of the quantity used in CR compounds throughout the rubber industry. Both powdered (100% active) and pre-dispersed (75% active) forms of ETU were used.
Synthesised accelerators were numbered SRM100 (SafeRubber Molecule 100) upwards for identification purposes. As with the compound numbers, the nomenclature carried no further implication.
Very small quantities of some of the accelerators were initially produced due to the complex nature of the synthesis and difficulties encountered with their production. In some cases only 5g were produced. With a minimum, mixable batch containing 100 g of CR, this only allowed one or two tests to establish the synthesised molecule’s potential effectiveness. Generally, 2.5phr of the SRMs was used for the first test batch as it was considered that this quantity would be enough to establish whether there was any effect on cure.
In cases where the rheometer results showed good cure efficiency, and there was enough accelerator available to produce a 200g batch, square sheets were moulded for physical testing to be performed.
SRM102 and SRM104 both showed reasonable cure characteristics, with SRM102 being the best, in that is cured quicker with a more distinct plateau than the slight marching modulus of SRM104. The rheometer trace for SRM102 is comparable to that of ETU. In fact at 2.5 phr it shows a quicker cure with a higher final torque and no marching modulus as displayed by the ETU compound. Further optimisation will take place in WP5.
Both SRM102 and SRM104 were chosen for scale up trials.
Environmental and Commercial Impact
Desk-based review of toxicity
The work focused on ETU as a reference for the new accelerator system under development.
Ethylene thiourea (ETU) has been used as accelerator in production of chloroprene rubber for decades. ETU is classified as CMR (Carcinogenic, Mutagenic or Toxic to Reproduction) and is thereby a substance that could identified as a Substance of Very High Concern and afterwards included in Annex XIV of REACH (Authorisation).
ETU is classified as CMR (Carcinogenic, Mutagenic or Toxic to Reproduction) and is thereby a substance that fulfils the criteria for being identified as Substance of Very High Concern (SVHC) and therefore as a future candidate to be included in Annex XIV of REACH (Authorisation). The main purpose with REACH is to ensure a high level of protection of human health and the environment.
Hazardous substances will be subject for Authorisation or Restrictions to ensure safe use.
Authorisation is a complex and heavy process. Relevant substances need to go through a process before they eventually are included in Annex XIV:
1. Identification of SVHC: Member States (MS) or ECHA on behalf of Commission include substance in Registry of Intentions
2. Submission of REACH Annex XV dossiers
3. Publication of Annex XV reports for comments
4. Development of Support Document (incl. responses to comments) by MS and ECHA
5. Agreement on identification by ECHA’s Member State Committee (if comments are received)
6. Inclusion in “Candidate list”
7. Immediate obligations following inclusion in the ‘candidate list’ for substances, mixtures or articles containing a “Candidate list” substance
The last substances were included in the “Candidate list” on 20.06.2011 and the list contains now 53 substances. The first 6 substances for Annex XIV was published 17.02.2011 and a corrigendum was issued the day after (Regulation (EC) No. 143/2011).
Article 67 in REACH states that “A substance on its own, in a mixture or in an article, for which Annex XVII contains a restriction shall not be manufactured, placed on the market or used unless it complies with the conditions of that restriction”. ETU use is currently restricted due to its classification as “toxic to reproduction category 1B”: Shall not be placed on the market, or used, as substances, as constituents of other substances, or, in mixtures, for supply to the general public when the individual concentration in the substance or mixture is equal to or greater than: either the relevant specific concentration limit specified in Part 3 of Annex VI to Regulation (EC) No 1272/2008, or the relevant concentration specified in Directive 1999/45/EC.
The key message is that ETU, the present used accelerator, has been classified as Reproductive toxicant category 1B with Hazard statement H360D – may damage the unborn child.
Being classified as CMR, ETU is a substance that could be identified as Substance of Very High Concern (SVHC) and a potential substance for inclusion in the Candidate List and later in Annex XIV of REACH (Authorisation).
Toxicity Testing
Understanding of the toxicological implications related to the introduction of new accelerator molecules is important. In line with the European regulations testing on animals should be avoided if possible, and toxicology databases can be a tool to be used to prevent unnecessary use of animal testing.
QSAR (quantitative structure activity relationship) has been an important study in the second year of the SafeRubber project. The outcome of a systematic QSAR study by UNIMIB based on 9 chemical scaffolds coupled with proposed molecules from rubber experts in the consortium was 15 molecules. Delivery Report 10 gave the toxicology testing results based on the use of QSAR tools. Twelve molecules are predicted to be safer than ETU.
Based on the on-going curing tests 2 molecules, SRM102 and SRM104 have given very promising results.
Both molecules have been predicted to be not mutagenic, not carcinogen and not skin sensitizer. However based on the laboratory trials SRM102 was chosen to be the best molecule for further development.
The selected sub-contractor LAB Research (Hungary) was acquired by CIT in June 2011 and the name was changed to CiToxLAB (Hungary). CiToxLAB has ca. 800 employees and has facilities in Hungary, France, Denmark and Canada.
QSAR (quantitative structure activity relationship) has been important in the SafeRubber project to screen for potential safe substances and to demonstrate that selected accelerator candidates are safer than ETU. The main purpose of QSAR was therefore not focused on REACH registration.
The outcome of a systematic QSAR study by UNIMIB was that 6 substances were predicted to be not mutagenic, not carcinogen and not skin sensitizer. Of these substances SRM 102 was selected based on its curing properties as accelerator for polychloroprene.
The REACH Regulation mentions (Q)SARs as a mean to provide data without doing animal testing. ECHA (European Chemical Agency) presented a paper in 2009 informing that results of QSAR may be used instead of testing when the following criteria are met:
1. Results are derived from a QSAR model whose scientific validity has been established
2. The substance falls within the application domain of the QSAR model
3. Results are adequate for the purpose of classification and labelling and/or risk assessment
4. Adequate and reliable documentation of the applied method is provided
5. In cases where there is uncertainty related to one or more information elements, QSAR results may still be used in the context of a Weight of Evidence approach
The main challenge with SRM 102 regarding QSAR is that this substance is a salt. Due to this the following actions were done:
• A screening of commercial tools able to handle QSAR predictions on disconnected structure was performed, but all the considered QSAR tools (ACDLAB - ACD/Tox Suite, Simulation plus, Accelrys - Discovery Studio TOPKAT) cannot support disconnected structures
• Queries on the databases of the OECD QSAR Toolbox were carried out in order to find disconnected structures to be used for read-across models. Few experimental data were available for salts and mainly for different endpoints, organisms and test conditions, making them not comparable and suitable to make predictions. Therefore, the set of disconnected structures in the QSAR toolbox was considered not significant in order to enhance reliable read-across models
• Relevant papers published in the scientific literature reporting QSAR models on disconnected structures were not found
In practice this means that our QSAR data should be used as supportive data for REACH registration. This means that the needed data in REACH Annex VII should be based on test data.
CiToxLAB, the selected sub-contractor for REACH testing, started testing end February 2013 based on a sample of SRM 102 made in Robinson Brothers Ltd. pilot plant. Tests are still ongoing.
The results so far have shown that SRM 102 is safe and the test data confirms QSAR predictions on toxicity (details are given in Delivery Report 14).
As part of Annex VII testing for REACH registration (volume 1-10 t/y) also Article 26 notification to ECHA is important. This work has also been sub-contracted to CiToxLAB.
Towards the end of the SafeRubber project ECHA announced in the “Registry of current SVHC intentions” that Sweden has filed a “notification of intention” of ETU (scope CMR) on the 16th April 2013. A likely consequence of this is that ETU will be included in the Candidate List during 2013.
QSAR validation
The entire set of ca. 52.000 molecules in the OECD QSAR Toolbox was considered initially by UNIMIB. By use of several “filters” the end result was the earlier mentioned 15 molecules.
A number of QSAR tools have been selected to get prediction of carcinogenicity, mutagenicity, developmental toxicity, acute toxicity, skin sensitisation, bioaccumulation and biodegradability.
QSAR tools designed to meet REACH requirements have been used. Such tools are OECD QSAR Toolbox, CAESAR, ToxTree, T.E.S.T. and Lazar.
Important criteria to meet when using QSAR are:
• Results are derived from a QSAR model whose scientific validity has been established
• The substance falls within the application domain of the QSAR model
• Results are adequate for the purpose of classification and/or risk assessment
• Adequate and reliable documentation of the applied method is provided
• In cases where there is uncertainty related to one or more information elements, QSAR results may still be used in the context of a “Weight of Evidence” approach.
The main purpose with QSAR in SafeRubber was to demonstrate that selected substances are safe. Validation and optimization of QSAR models have been an intrinsic part of the model construction in SafeRubber.
Environmental Impact Study
The initial plan was to make an evaluation of environmental impact of SRM102 throughout production, usage and disposal and compare by the impact of ETU, by use of ECETOC TRA (Targeted Risk Assessment) as offered by European Centre for Ecotoxicology and Toxicology of Chemicals (www.ecetoc.org). However, as the production of SRM102 has not been scaled up for commercial production yet and the toxicology testing by CiToxLAB has not been finalized the data are too few for a reliable TRA.
An ethic review was undertaken to minimize the use of animal testing by substituting QSAR results where applicable.
The scale-up was carried out at Robinson Brothers Ltd who are ISO14001 registered (Certificate No: EMS 57413) and before proceeding the established parameter check list was used as follows:
• Raw Material Assessment COSHH
• Safety Testing DSC
• Process Waste Analysis
• Quality Control Analytical Methods
• Energy Used
• Packaging Supply
• Disposal
• Equipment Required
• Transport Raw Materials
• Product to customer
• Odour
The following four steps have been defined for making a Life cycle assessment (LCA) according to ISO 14040 for replacement of ETU by SRM102 as accelerator in the production process of chloroprene rubber:
1. Goal and scope definition
• The objectives for doing the LCA comparison between SRM102 and ETU are:
• Better understanding of the value chain
• Insight on potential environmental impact risks
• Aligned economic, environmental and societal goals
• Engagement of value chain partners and stakeholders
• Identification of effective improvements
2. Inventory Analysis, including tracking material and energy flows in all relevant process steps:
This is an important part especially to collect the right information with respect to worker exposure and risk scenarios, and influence on the environment during accelerator production and vulcanization of chloroprene rubber, in addition to distribution, use and waste generation. Data needed to make an Impact assessment are:
• General physicochemical data of SRM102 and ETU (molecular weight, vapour pressure, water solubility, partition coefficients, in addition to results from biodegradability testing –
• In general, the routes of potential human exposure to SRM102 (and ETU) are inhalation, ingestion and dermal contact and the potential occupational exposure is greatest for workers involved in the manufacture of the accelerator and the manufacture of rubber and rubber products. Data needed for the worker exposure and risk scenarios (physical form, vapour pressure and operating temperatures. In addition information on ventilation, the use of respiratory and dermal protection to calculate inhalation exposure and dermal exposure are needed – (for SRM102 such data will be available when a scale up production has been designed, hopefully autumn 2013)
• Data on Energy inputs (track quantity, type and source of energy) – (data available after scaled up production of SRM102)
• Data on Production output including purity and eventually by-products – (data available after scaled up production of SRM102)
• Data on Emissions and Waste Outputs – (data available after scaled up production of SRM102)
3. Impact Assessment
For the Impact Assessment in this project it has been decided to use ECETOC TRA (Targeted Risk Assessment) as offered by European Centre for Ecotoxicology and Toxicology of Chemicals (www.ecetoc.org). However, as the production of SRM102 has not been scaled up for commercial production yet and the toxicology testing by CiToxLAB has not been finalized the data are too few for a reliable TRA.
An additional requirement was to try to reduce the heavy metal content (ZnO) of the compound and also to carryout leaching studies.
Reduction of heavy metal oxides
Polychloroprene rubber (CR) is commonly cross-linked by adding a combination of zinc oxide (5 parts per hundred rubber (phr)), magnesium oxide (4 phr) and an accelerator such as ethylene thiourea (ETU) or the new SafeRubber molecule, SRM102. Additional accelerators may be used to boost the cure rate.
According to the EPA “zinc ion can become available from zinc oxide through several mechanisms” and that “Zinc ion can reasonably be anticipated to be toxic to aquatic organisms”. One of the wider European benefits anticipated by the Safe Rubber Project is a reduction in the use of Zinc Oxide.
Two different experimental approaches were evaluated a) the replacement of ZnO by TiO2 and b) the replacement by a Multi-Functional Additive (MFA)
TiO2
A literature and patent search to identify previous work conducted with TiO2 was undertaken. It was identified that titanium dioxide has been used as filler in numerous elastomeric formulations but the use of this compound as a curing/vulcanising agent, in its own right (there are patents relating to titanium dioxide-containing complex organic molecules and titanium dioxide photocatalysts), is extremely limited (no patents were find to its use in polychloroprene formulations). However, a reference from this search was identified:
16/9/3. ‘Room temperature curable silicone rubber composition for use in e.g. electric appliance, composition comprises silicon polymer, surface treated silica, silane adhesive, silicone oil and titania type or zirconium-based curing catalyst’
Initial tests using TiO2 were performed at an early stage in the SafeRubber Project as part of a wider series of trials in conjunction with work investigating the curing mechanism of CR and also to provide data for QSAR computer modelling. When it was realised that TiO2, had the potential to replace zinc oxide (ZnO), it was decided to perform further work. The first series of tests were labelled SR134 to SR138. The five formulations were mixed and rheometer tested to establish their cure characteristics. The results are shown below. A measure of cure efficiency was established by deducting the minimum rheometer torque from the maximum.
It was established from this first series of tests that if ZnO is substituted by TiO2, when combined with the accelerators, HEXA and HVA enabled a good state of cure can be obtained. In fact the cure efficiency index is higher when TiO2 is used.
It was decided to investigate whether CR could be cured using TiO2 in place of ZnO using ETU as an accelerator with the addition of sulphur to the formulation and a good state of cure was obtained.
The optimum quantity of TiO2 was then investigated. Three compounds were formulated using 4 phr, 5 phr and 6 phr of TiO2, and these were based on the HEXA/HVA accelerator system.
Whilst the results of this brief test could be considered preliminary, it does appear that the TiO2 could be reduced to at least 4 phr without detrimental effect on the state of cure.
The final stage of work conducted during the Project was to investigate the effects of a reduction of sulphur and the combined metal oxides, in particular zinc oxide, or titanium dioxide, and magnesium oxide. Compounds SR245, SR246, SR247, SR248 and SR249 were used for this. The compounds were mixed and tested for their cure characteristics using a rheometer and for their physical properties.
The following conclusions can be observed from the results:
1. SRM102 gives a better cure than ETU, both in cure efficiency and physical properties.
2. A reduction of 1phr of each of the two metal oxides in the SRM cure compound does reduce the state of cure and physical properties; however the results are broadly in line with the ETU cured compound.
3. Reducing the sulphur content from 1 phr to 0.5 phr (active) at the same time as reducing the metal oxides does not greatly affect the cure.
4. If the sulphur is removed completely, reasonable cure efficiency is obtained but the physical properties drop considerably.
5. It is possible to replace zinc oxide with titanium dioxide in the curing system of polychloroprene rubber compounds.
6. It is possible to reduce the quantity of metal oxides from the current 5 phr ZnO and 4 phr MgO without greatly affecting the physical properties of the cured compound.
However there is still a great deal of further work which could be conducted, especially on the subject of ZnO replacement. The proportions of metal oxides and other components of the curing system all require optimisation and a full suite of physical testing needs to be performed once this is done, this work will continue post project.
Multi-Functional Additive (MFA)
A small amount of multi-functional additive (MFA) was synthesised to test whether it was possible to reduce the level of ZnO in a CR formulation. Mechanical tests have been completed with tensile testing results done after curing for 1.5 times T90, results including tensile modulus at various extensions, ultimate tensile strength and elongation at break were found. Similarly MDR cure results for a 15 minute test at 160 °C were found including maximum torque reached (MH), scorch time (TS1) and maximum cure rate.
The physical testing results show that the standard cure system produces similar results to SRM102 containing systems, in regard to both tensile results and cure results. It can also be seen that formulations 4 and 5 produce similar results to each other.
The conclusion from this small trial indicates that by incorporating MFA into a cure system it is possible to reduce the level of ZnO. It is recommended that partners who wish to use MFA carryout further optimisation work post project.
Leaching trials
In order to assess leaching of components of chloroprene rubber into rainwater over time, and thus assess environmental impact, a leaching experiment was set up. Samples of known weight (5g) of chloroprene rubber with different compositions were rinsed with rainwater and then placed into 25ml of rainwater and left for 12 weeks, with occasional stirring. This set-up aimed to mimic at a basic level the environment in a landfill or similar waste disposal site. After a 12-week period, the water samples were examined using UV-vis spectrophotometry to determine if any leaching of material had occurred.
The absorbance for all samples is very low. For the chloroprene gumstock sample cured with SRM102, no leaching of any components is observed after 12 weeks. For the other samples, absorbance values are all below 0.09 indicating exceptionally low concentrations of leached components. The lack of defined peaks also suggests the absorbance values observed are more likely to be down to light scattering by fines in the rainwater and machine error than due to absorbance by dissolved components.
Different samples of chloroprene rubber have been suspended in rainwater for 12 weeks and the water subsequently analysed by UV-visible spectrophotometry. Data obtained suggest that leaching of the component chemicals into the rainwater over the 12 week period is minimal and does not provide an environmental risk.
UV testing
UV testing was carried out according to EOTA (European Organisation for Technical Approvals) Technical Report specifies exposure procedures for artificial weathering using EN ISO 4892. The radiant exposure range in Northern Europe is typically 128MJ/m² per year.
The laboratory light source type 1 (UV-A 340 nm peak) was used where radiant emission below 400 nm makes up at least 80% of its total light output and where radiant emission below 300nm is less than 2% of its total light output, with a spectral irradiance of (45 ± 10) W/m2 in the bandpass of 300 nm to 400 nm.
The simulation exposure conditions were for a moderate climate in Europe. The testing cycle was UV radiation for 4 hrs at 60°C then a condensation cycle for 4 hrs at 50°C then repeated. The lamps run at ~40W/m² (300-400nm). Testing was for 500 hours, this is equivalent to half a year in the Arizona desert.
The test samples were black filled and gumstock moulded compounds:
• SR141, ETU gumstock
• SR179, SRM102 gumstock
Tensile testing before and after exposure was used to determine the effects of UV exposure and the results show that there is no difference between compounds made with ETU and SRM102 and both show that there is negligible degradation due to UV radiation.
Conclusion
Chloroprene compounds containing SRM 102 have been tested in a wide array of industrial processing techniques in five factories in three different countries. In every case SRM102 has equaled or exceeded compound containing ETU. It has proved itself to be a viable replacement for ETU as an accelerator Deliverable report D18).
4. Interpretation
The interpretation including limits, gaps, recommendations and conclusions will be made after the TRA has been finalized. However, Chloroprene compounds containing SRM 102 have been tested in a wide array of industrial processing techniques in five factories in three different countries. In every case SRM102 has equalled or exceeded compound containing ETU. It has proved itself to be a viable replacement for ETU as an accelerator Deliverable report D18). Some findings from these tests relevant for environmental impact are:
• Along the experience with the vulcanization trials, the ability to flow and fill easily the mould has been found to be much better in the CR rubbers made with SRM102 than in the standard ones. This could arise in a reduction of the mass of rubber needed to vulcanize a piece which in turn will reduce the scrap.
• The cure cycle could easily be reduced through optimisation of the accelerator system
• ZnO is very toxic to aquatic organism. It is possible to reduce the level of ZnO, by using either TiO2 or MFA’s especially when SRM102 is used as the accelerator; however further trials post project will be needed to investigate other effects such as ageing and oil resistance etc.
• Rubber samples have been placed into rainwater. After 12 weeks samples look similar. However, algae growth on samples produced by use of SRM102 suggests they’re not toxic!
• There is no difference in the UV degradability between ETU and SRM102 compounds.
Commercial Feasibility study
A commercial feasibility study will be undertaken throughout the project to ensure the development of the replacement molecule is commercially viable (is cost competitive to state-of-the-art materials i.e. the manufacturing cost of a final rubber compound is less than 105% of the equivalent compound made using ETU). The study will be directly relevant to the lab scale development in WPs 1&2 and will feed into the scale up activities in WP4 and industrial validation in WP5.
The trials carried out in WP 5.2 (rubber manufacturing trials) and reported in WP 5.3 have proven that the SRM102 is a commercially feasible replacement for ETU.
Robinson Brothers Ltd is currently unable to estimate the definitive price of SRM102 until it produces the product on an industrial scale (it has only produced 44 kg so far). However all of the trials carried out indicate that the proportion of ETU and SRM102 used in the formulas for all the end users is very similar i.e. around 1% , therefore the current cost estimates of the accelerator SRM102 will not increase the manufacturing cost of a final rubber compound more than 105% of the equivalent compound made using ETU.
For a more in-depth report showing tables and graphs, please refer to the attached PDF.
Deviations & Corrective Actions:
None
Work package 4: Scale Up of Replacement Accelerator
Development of suitable synthesis route and process conditions
According to the preliminary curing tests two possible new accelerators SRM102 and SRM104 were selected for Kilo Lab-scale process.
It was found that the syntheses of SRM102 and SRM104 accelerators were complicated and accompanied by the formation of several by-products. The formation of undesired products is due to similar pKa values and similar reactivity of both free amino groups in the starting compound. As a consequence a new, reproducible procedure for the preparation of the target products SRM102 and SRM104 on a 1 kg scale as one batch syntheses was developed. The reaction conditions, reagents, solvents, temperature and reaction time were been optimised.
All intermediates, by-products and final products were analysed by 1H NMR spectroscopy. The purity of the key intermediate and products SRM102 and SRM104 was analysed by HPLC-MS technique However, under standard conditions, HPLC-MS analysis is not very reliable. There is the possibility that the analytes can interact with sorbent or mobile phase. The greatest problem is pH of the eluent: exchange of eluent to pH>9 (suitable for solubility of betaine) will result in even lower sensitivity. The use of high pH values and various salt additives will result in extremely low lifetimes of HPLC standard columns.
The new, scalable and reproducible method for the preparation of the key intermediate betaine was elaborated. All possible alternative methods and by-products were examined and analysed by 1H-NMR. The optimal conditions were found and the intermediate betaine was obtained in 80-95% yield with the purity 99% by NMR.
With betaine as the key intermediate, the optimal reaction conditions of the target accelerators SRM102 and SRM104 have been determined and reported.
The purity of synthesised products: betaine, SRM102 and SRM104 was analysed using acid-base titration analysis. The acid-base titration is one method for the quality control of betaine and final products SRM102 and SRM104. In this case side interactions of target products and used materials for titration were not observed.
Scale up process on the Kilo Lab Scale
The new accelerator SRM102 was prepared using the synthetic procedure outlined in Deliverable 15 at Kilo Lab scale i.e. 1kg batch trial in quantitative (>99%) yield with the purity 97.5% combination of 1H-NMR (impurities) and acid-base titration (assay) and the new accelerator SRM104 was obtained in 1kg batch trial in quantitative (>99%) yield with the purity 99.1% combination of 1H-NMR (impurities) and acid-base titration (assay).
The bulk synthesis compounds were analysed and the results compared to the compounds synthesised at small scale using both 1H-NMR NMR and acid-base titration. Analysis showed the compounds to appear chemically identical.
To ensure curing efficiency was also maintained, bulk SRM102 and SRM104 samples were used in chloroprene curing experiments and the results compared to small-scale-synthesised SRM102 and SRM104 chloroprene cure curves. It is clear from the cure curves that the Kilo Lab produced samples of SRM102 and SRM104 behave in the same way as the small-scale-produced samples do and so scale up to pilot scale can now be evaluated.
Novel chloroprene curing agents SRM102 and SRM104 have been successfully produced at Kilo Lab scale and shown to behave in the same way as samples produced at small scale in the curing of chloroprene rubber.
Scale up process on Pilot Plant Scale
Kilo quantities of SRM102 and SRM104 have been made by GSL and a synthesis route determined and reported in D15. RBL are currently assessing the proposed synthesis routes. Further studies carried out by the research Partners on the two potential candidates has shown that SRM102 shows improved rheological and physico-mechanical properties over SRM104. As a consequence, the Consortium has agreed that scale-up should concentrate on SRM102 only.
A series of validation trials was carried out in the laboratory and then kilo scale in order to ascertain the robustness of the process and to obtain the critical Health and Safety data for the process before proceeding to Pilot Plant scale. At the validation stage, modification to the process may be required to improve and optimise the process and yield. It should be noted that process development and optimisation may alter the final product, e.g. purity, and this may have an effect on the efficacy of the accelerator molecule compared to the research developed/synthesised equivalent. This will need to be assessed by the research Partners.
Quantities of SRM102 manufactured at the validation stages will be made available to Clwyd Compounders and Mixer to allow them to carry out their own compounding optimisation trials whilst the pilot plant quantities are being manufactured.
RBL plan to carry out validation trials in June/July and scale up to Pilot Plant in mid-August/September. The time-scale provided assumes that the process provided will not require any changes. Any process development and optimisation will increase this time-frame. Validation trials on laboratory and kilo-scale will need to be carried out to determine the robustness of the process and to provide critical health and safety data required before the process can be carried forward to Pilot Plant scale. Any process improvements made to the process may alter the final product and therefore its efficacy may need to be reassessed.
RBL plan to start process validation trials in June and provided no process development is required, Pilot Plant scale-up is scheduled from mid-August/September. This time-frame will increase if process development and optimisation is required.
The aim of this subtask was to determine methods for the reproducible synthesis and scale-up of production of SRM102 and SRM104, going through kilo- and pilot-scale preparation. In order to achieve this, it was first necessary to determine a suitable synthesis route for scale-up, based on those outlined previously in WP2. Before the scale-up could be started it was necessary to ensure that the Robinson Brothers parameter check list was followed:
• Raw Material Assessment COSHH
• Safety Testing DSC
• Process Waste Analysis
• Quality Control Analytical Methods
• Energy Used
• Odours
• Packaging Supply
• Disposal
• Equipment Required
• Transport Raw Materials
• Product to customer
• Odour
During the lab-scale production of SRM102 and SRM104 the synthesis routes were modified slightly taking into account the above parameters and the synthesis was determined to be quite complicated and careful control was required in order to avoid making by-products. These synthesis routes were used as a starting point for the kilo-scale synthesis of the new accelerator molecules, employing 1H NMR spectroscopy and HPLC as analytical techniques to support synthesis optimisation.
It had already been observed in WP2 that successful production of the accelerator molecules depended heavily on the acidity of the reaction mixture, due to the varying pKa values of the reaction intermediates. If the acidity of the reaction mixtures was not carefully controlled, the reaction mixtures could decompose or undergo undesired reactions. By studying the various pKa values of the reaction intermediates and products, a synthesis route was determined that first produced a betaine intermediate, which was common to both SRM102 and SRM104. This betaine could then be further reacted with the desired diamine to produce the accelerator compounds.
With this knowledge, the synthesis route at kilo-scale was undertaken and optimised to yield SRM102 and SRM104 in 80 – 95% yields. The next important stage in the scale-up procedure is the development of a suitable assay method to determine purity. The most widely used purity assessment method is HPLC, due to its reproducibility and speed of analysis. This method was therefore selected for purity determination of these accelerator compounds. However, as SRM102 and SRM104 are salts, determination using HPLC is not straightforward. This is because salts dissociate when dissolved and so it is difficult to determine the purity of the compounds using one analysis method alone.
Initially HPLC-MS was used, but the results from this analysis method were not reliable or reproducible. It was thought that this may be due to the analytes interacting and reacting with the HPLC column material or decomposing during analysis. Another major issue was the need to run analyses at pH>9 to ensure solubility of the components. This leads to short column lifetimes and reduced sensitivity.
To overcome this, an acid-base titration technique was first determined to assess purity as a replacement for HPLC analysis. However, once again, the sensitivity of this technique was low and the variation in results was large. In addition, analysis times were too long and laborious to be used at pilot- and production-scale. Therefore, analysis by HPLC was pursued again, this time not employing mass spectrometry. After further optimisation of the analysis methods, it was decided to develop methods for the analysis of the anion and the cation separately. This was significantly easier than trying to devise a combined analysis method and the purity of the SRM102 and SRM104 compounds could therefore be assessed successfully.
The HPLC traces obtained for the SRM102 and SRM104 compounds showed that the individual dissociated anion and cation were both around 97% pure after the first kilo-scale synthesis by GSL. The synthesis and analysis methods were therefore handed over to RBL, where they were carried out at kilo-scale in their kilo-lab production facilities. Using HPLC and TGA analyses, the resulting products were compared to those produced by GSL and found to be similar, showing the synthesis procedure was reproducible and produced the desired products in high yield and purity. The compounds were also found to act as curing agents in the same way as the products produced at small scale.
Finally, the synthesis method was developed into a pilot-scale production method. It was decided on the basis of data on toxicology, safety and product quality that only the large-scale production of SRM102 would be undertaken. This was achieved at pilot-scale, with 44kg of SRM102 being produced and analysed.
Work package 5: Rubber Curing, Component Manufacture and Industrial Validation
Rubber Compounding Trials
The SRM102 production molecule was supplied to MGN, Mixer, Clwyd and MaTRI.
MGN, Mixer, Clwyd each carried out rubber compounding optimising trials to find the best formulation(s) for their applications. Further more details results can be found in 5.3
MaTRI carried out further generic optimisation trials for dissemination purposes.
Rubber Component Manufacturing Trials
Production of rubber components is performed using a great variety of manufacturing techniques and it was considered by the Safe Rubber Consortium to be important to utilise as many methods as possible during the industrial validation trials.
The process starts with mixing, or compounding the elastomer with various ingredients. These include curatives, anti-degredants, processinng aids, fillers and other functional additives. Mixing is carried out either on an open two-roll mill or in an internal mixer..
After mixing the compound must be shaped by methods such as moulding, extrusion or calandering. Once formed into the desired shape the component must be cured, alternatively refered to as crosslinking or vulcanisation.
An industrial trial programme was designed to test SRM102, the newly developed acellerator, as a viable alternative to ETU. Because SRM 102 is currently made in a powder form and used in small quantities of less than two parts per hundred rubber, it is unlikely that it will affect the physical processing behaviour of the compound, for instance mixing and extrusion speeds, dimensional charateristics, heat build up etc. However its cure characteristics could initiate several problems.
These include:
• Premature cure, or scorching, during the mixing process. In its mildest form scorching can take the form of an increase in viscosity and small lumps of cured rubber developing throughout the compound. With the higher temperatures found in an internal mixer, the whole compound can start to cure which can be detected by rapidly rising torque readings
• If scorch is initiated by the heat in an extruder barrel, the result is a rough extrudate
• Scorch during calendering will cause a rough surface finish
• A slow curing extrudate is liable to sag and become misshapen during an autoclave cure
• Porosity in the rubber can occur in a slow curing continuous cured compound
• Mouldings can become distorted if the cure state is incorrect
The Consortium developed a test matrix, which included the types of process, types of product and partners with the capability to perform the trials. It was evident that the Consortium did not have the capability to perform all of the necessary processes. Two companies outside the Consortium offered to provide their assistance with the trials on the basis that the aims of the Project offer great potential to improve safety in the European rubber industry in the future.
Process Application Performing
Extrusion Autoclave cure Round section Nufox Rubber*
Garvey die section Mixer
Continuous cure Round section Nufox Rubber*
Polychloroprene cable sheathing Mixer
Moulding Compression Thin sections MGN
Thick sections MGN
Rubber to metal bonding MGN
Injection moulding Various BD Technical Polymer*
*Non-Consortium Member
Notes:
• As both compression moulding and injection moulding trials were successful, transfer moulding was deemed by the Consortium rubber experts to be un-necessary because it is a hybrid of the two processes.
• Calendering was not trialled on the grounds that it would require a large quantity of compound along with lengthy machine set up times and as such was too expensive. Furthermore the rubber experts in the Consortium agreed that if the extrusion trials were successful, then there would be no reason why calendering would not also be successful.
• Polychloroprene sheathing trials were performed using a specially developed compound to meet industry specifications. This utilised a cross-head extruder.
• Two types of continuous cure were trialled; infra-red and salt-bath.
The manufacturing trials were performed or overseen by four Consortium partners; MaTRI, Mixer, Clwyd and MGN. This report describes the trials by each partner.
Full Analysis of Industrial Scale Validation
As previously mentioned, the Safe Rubber Consortium did not have the capability to perform the full range of trials that was required to validate SRM102 in an industrial environment. MaTRI located two companies outside the Consortium who offered to assist the Project for no fee.
MaTRI provided all partners with guidance and health and safety information. The technical guidance was taken from work performed with SRM102 and SRM104 earlier in the project. The health and safety information was based on work by GSL. The guidance notes are can be found in the appendix od D18.
Injection Moulding
An injection moulding machine consists of a cylindrical injection barrel with a ram or screw inside it. Rubber compound is forced towards a nozzle at its end which is connected to the top half of a closed mould. Within the mould are gates and runners leading to the mould cavity itself. Cure temperatures used in injection moulding can be much higher than those used for compression or transfer moulding enabling faster cure cycles.
Injection moulding trials were performed by BD Technical Polymer, which is a member of the BRPPA. The compound used for the injection moulding trials was the same as Trial Compound 1 used for the continuous cure trials.
A semi-spherical, hollow, parabola-shaped product was chosen for the trials. Its large diameter was approximately 130mm with a height of 60mm. The small end was recessed at approximately 42mm diameter and contained a 10mm hole. The wall thickness was 4mm. The net weight was 175 grams per product. Figure 6 shows the details of the product.
An REP RT9 vertical injection moulding machine was used. The cure temperature was set at 160°C for all trials as being representative of industry practice and also within the operating temperature range of SRM102.
The machine was thoroughly flushed out with the test compound before series of mouldings was produced using a single cavity mould.. The first pieces were moulded for 25 minutes and appeared to be fully cured. State of cure was assessed by general feel and an indentation resistance and recovery test. No moulding defects or abnormalities were identified in the test pieces.
Similar results were obtained after reducing the cycle time to 20 minutes and then to 15 minutes. Moulding for 10 minutes produced samples which were not considered to be fully cured. Further samples were cured using the 15 minute cycle to confirm the earlier results.
The injection moulding trial was considered to be a success. No moulding defects were observed, the compound cured after the mould was fully loaded and there was no indication of premature cure (scorch) occurring. The 15 minute cure cycle could easily be reduced either through optimisation of the accelerator system or by using increased temperatures. The products obtained were of equal quality to products made using traditional ETU accelerated compound.
Continuous cure (non-cable compound)
Two types of continuous cure or vulcanisation (CV) were trialled with compound containing SRM102; salt-bath and infra-red. Both methods use a normal cold feed extruder, the main difference being the method by which heat is provided to the un-cured profile.
Salt bath CV is a commonly used liquid curing method for extrudates. CV is frequently chosen for producing products such as tubing, hoses and weather stripping. Salt baths are relatively short-length curing units because salt has good heat exchange properties and can be used at high temperatures of up to 260°C. Salt does not cause surface oxidation, and is easy to clean off the finished product using water.
Infra-red (IR) CV lines use a row of IR heaters to cure the extrudate and tend to be longer because heat flow into the rubber compound is slower when compared to a salt bath.
In both processes it is essential that the compound starts to cure enough to maintain its shape before distortion due to melting occurs. Another consideration is the formation of porosity which can be caused by the formation of volatiles before cure has taken place. It can also be caused by moisture in the compound turning to steam.
Continuous cure trials were performed by Nufox Rubber on both its infra-red and the salt-bath extrusion lines. Autoclave cure trials were also carried out by Nufox. The trials were supervised by MatRI and Clwyd. Clwyd provided the compound for the trials.
The compound was extruded through a 10mm circular die and various parameters were changed during the course of the trials until optimum conditions were obtained.
Whilst it cannot be considered as continuous cure, samples of extrudate were taken from the extruder line and cured in an autoclave. This process consists of placing the un-cured extrudate on a talc covered, aluminium tray in an atmosphere of pressurised steam. The pressure used was for the trials was 4 bar which equates to 152°C.
Mixer is a manufacturer of rubber and thermoplastic compounds for the insulation, sheathing and filling of flexible cables, rubber cables, power cables and flame-retardant cables. This is a specialist market due to the use of highly specified material with high volumes and low margins, making it different from the general purpose rubber goods industry.
Mixer’s task in the Saferubber project was to validate the use of the new accelerator SRM102, in an industrial manufacturing envoirment and test the substitution of ETU in rubber cables. The overall aim being to reduce the toxicity caused by fumes from ETU during manufacture without losing technical performance.
Mixer’s first task in this part of the SafeRubber project was to validate the use of the new accelerator in industrial manufacturing and to test the substitution of ETU in rubber cables in order to reduce toxicity issues during production without losing technical performance.
SRM102 was previously chosen as the best fit to replace ETU for both rheometric and mechanical properties. This accelerator will be tested as a direct replacement in the compound formulation and as a complete replacement of the whole crosslinking package.
Rheometric, mechanical and ageing properties will be investigated in order to understand if this chemical will be suitable to produce heavy duty, high temperature, oil resistant cables used worldwide for offshore and mining applications.
Five batches were produced and tested in order to check the reliability of the compounding procedure. Just one batch was discharged because it was not fully homogeneous. The remaining four batches were mixed in order to obtain a homogenous amount of compound to be divided and modified with the addition of different crosslinking packages.
During mixing it was quite clear that in the SRM102 sample, lumps of the accelerator were present affecting the dispersion of the chemical.
However the test on the masterbatch was completed in order to show the failure produced by the lumps. Results show lower mechanical and rheometric properties which are caused by poor dispersion of SRM102.
The next set of industrial trials were designed to test the extrusion properties and manufacturing conditions in order to understand the critical features during processing. A standard, currently used, formulation was chosen for the trials: 5GM3 which meets DIN VDE 0207 teil 21.
It was clear that the crosslinking speed of the SRM102 compound is lower (higher t90). For this reason, two new test sets were prepared in order to test the SRM102 concentration effect on the properties, and the best balance of the accelerators in presence of SRM102.
SRM 102 concentration acts as a booster for the vulcanization process increasing the speed and the crosslinking density and mechanical properties are slightly improved when more than 2 phr of SRM102 is used.
In order to understand how SRM102 works together with other accelerators UNIMIB prepared a design of experiments programme. The influence of secondary accelerators was investigated.
TMTM increases compounding and handling safety (increasing ts2) and productivity (reducing t90).
SRM102 and DPG act as standard accelerators reducing ts2, and increasing the crosslinking density.
Sulphur acts similarly to DPG and SRM102 but without reducing t90. Sulphur appears to be the best solution to give increased MH, and for this reason even hardness.
All the experiments performed by Mixer show that SRM102 could be an effective replacement of ETU in polychoroprene based compound.
Crosslinking density, hardness, scorch safety, mechanical properties and more importantly, ageing resistance is very close to what could be obtained with ETU. Crosslinking speed is slower but from the data obtained from UNIMIB’s DoE it is clear that TMTM and DPG are useful additives for tuning the performance of SRM102 for a complete and successful substitution of ETU in heavy duty service cable compounds.
Two compounds were used to compare the effectiveness of SRM102 compared with ETU.
The trial compounds were extruded through an extruder. The extrusion temperatures and pressures were virtually identical for both compounds.
The extruder was fitted with a Garvey die to test the extrusion properties because it was felt that this would highlight any problems such as scorch. It was seen that there is little or no difference between the two compounds.
In conclusion, it can be stated that there is no discernible difference between the processing characteristics of compound containing ETU or SRM102. Both extruded in an identical manner with no problems or issues.
Clwyd Compounders Ltd. was required to carry out industrial evaluation studies on the use of SafeRubber accelerators SRM102 and SRM104 in polychloroprene rubber compounds.
20kg of unfilled polychloroprene masterbatch was mixed and sheeted on a production scale two-roll mill. SRM102 and SRM104 accelerators were added to aliquots of masterbatch on a two-roll laboratory mill at loadings from 0.5phr to 2.5phr blended well and sheeted off.
The rheology of each compound was tested at 185°C for 6 minutes. Cure conditions were based on the cure curves obtained. 2mm thick test sheets and small (6.3mm x 13mm) compression set buttons were moulded of each compound for 15 minutes at 165°C.
Conclusions - cure efficiency initially increases with increased loading of SRM102 but then plateaus at 1.5 phr SRM102. However, cure speed continues to increase at loading of more than 1.5 phr SRM102, increasing scorch risk. Optimal physical, aging and compression set properties were seen at 1.5 phr SRM102. The cure efficiency of SRM104 containing compounds gradually increases with loads up to 2.5 phr SRM104 but so does cure speed. Optimal physical properties, aging and compression set properties are seen at 1.5 phr SRM104.
A study was done to investigate the use SRM102 and SRM104 in commercial polychloroprene compounds by substituting a standard accelerator system for SafeRubber accelerators in a commercial black-filled compound. The resulting compounds were tested against BS 2752 and DTD 5514 standards and results compared against previously obtained results in an ETU containing benchmark compound.
Two compounds 30M3B023 & 30M3B024 were mixed in laboratory internal mixer, then slabbed, re-milled and sheeted on a two-roll laboratory mill. The rheology of each compound was tested at 165°C for 30 minutes. The rheology of each test compound was tested for 30 minutes at moulding temperature (165°C). A 195 x 195 x 2mm test sheet was moulded of each formulation for specification testing against both the BS2752 and DTD5514 specifications.
In conclusion both general purpose and high-quality SRM102 compounds show good basic properties. The high-quality CR compound is within reach of the challenging DTD 5514 specification requirement, which has traditionally required ETU containing compounds to achieve it. Although SRM102 & SRM104-containing general purpose CR compounds show low elongation compared to other accelerator systems, tensile strength and modulus at 100% elongation are higher. Compression set and fluid resistance testing showed comparable results to the standard accelerator systems, suggesting that a good state of cure is achieved in a general-purpose CR compound using 1 phr of SafeRubber accelerators. The short t05 time may lead to shelf life issues, which suggests that shelf life studies on SRM102-accelerated compounds may require consideration.
MGN’s role in Work Package 5 was to validate the use of SRM102 in moulding compounds an industrial environment, specifically compression moulding of various sized parts and rubber to metal bonding. MGN does not possess any injection or transfer moulds. MGN selected some pieces currently manufactured by the company with chloroprene rubber in production moulds, in order to test the whole range of applications described earlier.
The work to be done is to manufacture all these parts, substituting the current ETU containing compound used by MGN for a new compound, having the same formula (or as similar as possible), but substituting ETU with the new accelerator developed in the SafeRubber Project, SRM102.
The industrial validation of the suitability of the new accelerator for its use in the vulcanization of CR consists on the comparison of several significant parameters, with those obtained with the current CR used by MGN:
MGN uses basically two different CR blends, so before vulcanizing the chosen references, a previous step had to be done, to substitute the ETU by SRM102 in those two blends. During this step, various combinations in the formulations where tried to introduce SRM102 in MGN’s CR blends, having as reference original MGN CR blends to compare the results obtained.
The study was performed in two steps:
• Step one: a formulation was chosen and some trials on real moulds were made.
• Step two: Some more versions of the formula were made, trying to improve the vulcanization performance and one was chosen
General conclusions from the vulcanization curve:
• Suitable viscosity in all of them
• Enough scorch time in all of them
• Slower vulcanization rate when using SRM102 at tested amounts (up to 1.8 phr; if more used, insufficient scorch time would appear)
• Similar level of vulcanization than the original one can be achieved depending on the SRM102 phr
• The use of TMTM instead of TMTD doesn’t improve significantly the vulcanization
MGN has optimized its two CR formulas to be used substituting the ETU by the new accelerator SRM102, and has tried them in the moulds proposed in order to check all the technologies and applications described in Table 27.
The new CR formulas with SRM102 have been found to work properly in all the tested technologies and applications.
In a first set of trials, it seemed that the use of the new rubbers could arise in a loss of productivity, as the vulcanization curves suggested that a little extra time should be used for vulcanization. However, a second set of trials was successfully conducted to check and prove that the new rubbers could be used in the tested moulds with the standard vulcanization time, so the use of these doesn’t imply any productivity loss.
In addition, some characteristics of the new rubbers have been found to improve the ones of the standard formulas:
• New formulas have higher tensile strength than the standard ones, so a characteristic as important as the tension at break in a piece like the guiding part tested, results to be higher if the piece is manufactured with the new CR.
• Along the experience with the vulcanization trials, the ability to flow and fill easily the mould has been found to be much better in the CR rubbers made with SRM102 than in the standard ones. This could arise in a reduction of the mass of rubber needed to vulcanize a piece
As a general conclusion, it can be said that the industrial validation of the new CR rubbers made with SRM102 has been successfully achieved, for all the technologies and applications tested by MGN.
Chloroprene compounds containing SRM 102 have been tested in a wide array of industrial processing techniques in five factories in three different countries. In every case SRM102 has equalled or exceeded compound containing ETU. It has proved itself to be a viable replacement for ETU as an accelerator.
Commercialisation
RBL is unable to estimate the definitive price of SRM102 until it produces the product on an industrial scale (it has only produced 44 kg so far). At the moment is not possible to compare the current cost of MGN’s standard formulas with the one of the new formulas, although the variation should be very small, due to the small proportion of ETU and SRM102 used in the formulas.
Potential Impact:
The SafeRubber project has provided the SME members with an IP protectable method (Trade Secret) to manufacture chloroprene rubber, which will enable them to effectively compete with large enterprises and Asian importers. The European market for chloroprene rubber is over €210 million per annum10 and the Global market ~€650 million10 per annum. Assuming a modest market penetration of 10% by 2017 (five year post project), this equates to increased sales (displace imports) of €45.3 million cumulatively for the SME members. This extra €41.5 M turnover will help safeguard or create 453 jobs12 within Europe.
In addition to this, it has been conservatively estimated that a market penetration of 3% by 2017 outside the EU (exports) equates to sales of €36.4 million creating or safeguarding an extra 364 jobs within Europe. This project also supported the SMEs who supply the chemicals and processes to distribute the product developed to the rubber industry.
A replacement that is safer than ETU has been developed i.e. SRM 102 which has proven to be a more environmentally friendly process, which in turn leads to inherently fewer hazards and less arduous health and safety requirements. The accelerator has also been successful in processes i.e. epichlorohydrin rubber (ECO).
SafeRubber has produced a more environmentally friendly process, which in turn will lead to inherently fewer hazards and less arduous health and safety requirements.
During the moulding trials it was proved that compound made with the new accelerator also had better flow properties, this should result in less scrap being produced, hence less landfill. This has been industrially validated by all the end user beneficiaries with comments that this has equalled if not exceeded that curing properties of ETU.
The process has developed a platform technology that has been proven that it can be applied to other processes for example, epichlorohydrin rubber (ECO).
Additional trials have also shown that it is possible to replace/reduce the ZnO which has a significant environmental benefit by using a multi-functional additive or TiO2 thus the environmental impact can also be reduced through significantly lowering the quantities of MgO and ZnO required in the process – conserving raw materials and reducing the quantities of hazardous effluent.
The project has been successful and now means that there is an accelerator which is suitable for the production of poly chloroprene and is now not gender restrictive.
Considerable interest has been shown in the new molecule which should benefit both the consortium and the European rubber industry
List of Websites:
www.saferubber.eu
Coordinator - Girolamo Dagostino - g.dagostino@assocomaplast.org
Technical Manager - David Cartlidge - david.cartlidge@pera.com
Dissemination Manager - Geert Scheys - gscheys@essenscia.be
