VOLUME 47 2012 ISSUE 1 www.uctm.edu ISSN 1311-7629 (print)

ISSN 1314-3859 (on line)

JOURNAL

OF THE UNIVERSITY OF CHEMICAL TECHNOLOGY AND METALLURGY

SOFIA Journal0B of the University of Chemical Technology and Metallurgy UCTM has been publishing the journal called Annual of the University of Chemical Technology in Bulgarian language since 1954. Thirty four volumes have been issued. It has been transformed into a Journal of the University of Chemical Technology and Metallurgy published in English since 2000 (volume 35). JUCTM is published quarterly.

Honorary Editor R. Dimitrov

Editor-in-Chief1B B.5B Koumanova

University2B of Chemical Technology and Metallurgy,

84B Kl. Ohridski blvd., 1756 Sofia, Bulgaria

Tel:3B (+ 359 2) 8163 302, E-mail: [email protected]

Editorial7B Board

S. J. Allen V. Meško Queens University of Belfast, UK International Balkan University, Skopje, Macedonia D. Angelova L. Mörl UCTM, Bulgaria University “Otto-von-Guericke”, Magdeburg, Germany M. Bojinov B. Nath UCTM, Bulgaria European Centre for Pollution Research, London, UK J. Carda K.D.P. Nigam University Jaume I, Castellon, Spain Indian Institute of Technology, Delhi, India G. Casamatta L. Petrov Institut National Polytechnique de Toulouse, France Bulgarian Academy of Sciences, Bulgaria G. Cholakov S. Ramsuroop UCTM, Bulgaria Durban University of Technology, South Africa V. Dimitrov D.C. Shallcross Bulgarian Academy of Sciences, Bulgaria The University of Melbourne, Australia N. Dishovsky M. Simeonova UCTM, Bulgaria UCTM, Bulgaria S.J.C. Feyo de Azevedo V. Stefanova Universidade do Porto, Portugal UCTM, Bulgaria M. Jitaru D. Stoilova University “Babeş -Bolyai”, Cluj-Napoca, Romania Bulgarian Academy of Sciences, Bulgaria S. Kaltcheva I. Turunen UCTM, Bulgaria Lappeenranta University of Technology, Finland F. Keil S. Vassileva Hamburg University of Technology, Germany UCTM, Bulgaria T. Konstantinova V. Vassilev UCTM, Bulgaria UCTM, Bulgaria M. Kucharski S. Veleva AGH University of Science and Technology, Krakow, Poland UCTM, Bulgaria A. Mavrova L.Vezenkov UCTM, Bulgaria UCTM, Bulgaria D. Mehandjiev Ž. Živković Bulgarian Academy of Sciences, Bulgaria University of Belgrade, Technical Faculty, Bor, Serbia

Technical secretary: S. Georgieva

The annual subscription (for 4 issues) is 200 €, including postage, handling and packaging charge.

©UCTM, 2005. All rights reserved. Reproduction in part or whole without permission is strictly prohibited. Prepress and design: A. Staneva; Printing: S Print Ltd., Sofia

The articles of this journal are indexed and abstracted in EBSCO, Chemical Abstracts and Referativnii Journal Himii (VINITI). (www. http://www.uctm.edu) TABLE OF CONTENTS

REVIEW Addvanced approaches of industrial informatics in the chemical industry I. Batchkova...... 3

CHEMICAL TECHNOLOGY

Energy savings in alkyd-resin and aqueous-emulsion processes using a heat-integrated cycle Mohammad Al-Shannag...... 31

Thermal ageing of different kinds of fiber materials: A Kinetic study R. Boeva, G. Radeva, P. Hinkov, E. Hinkov...... 37

Forest fertilizer properties of the bottom ash and fly ash from a large-sized (115 MW) industrial power plant incinerating wood-based biomass residues H. Nurmesniemi, K. Manskinen, R. Pöykiö, O. Dahl...... 43

BIOTECHNOLOGY

Technology for production of fluoride enriched beer G. Yonkova, A. Surleva, T. Ginova-Stoyanova...... 53

Effect of Saccharomyces Boulardii yeasts addition on the taste and aromatic properties of kefir G. Ivanova, M. Momchilova, N. Rumyan, A. Atanasova, N. Georgieva...... 59

Pretreatments effect of sorghum (Bicolor (L.) Moench) and millet (Pennisetum Glaucum) flours on the in vitro starch digestibility Rachid Souilah, Badreddine Belhadi, Nadia Boudries, Djaafar Djabali, Boubekeur Nadjemi...... 63

CHEMISTRY AND SYSTEM ENGINEERING

Removal of cobalt and nickel from wastewater by using Jordan low-cost zeolite and bentonite Reyad A. Al Dwairi, Aiman E. Al-Rawajfeh...... 69

Advanced software architecture of an automatic vehicle number plate recognition system A. Atanassov...... 77 Comparative analysis of case based reasoning software frameworks jCOLIBRI and myCBR A. Atanassov, L. Antonov...... 83

CHEMISTRY

Determination of the oxygen stoichiometry of Y Ba Cu O superconductors 1 2 3 y with different oxygen content St. Georgieva, T. Nedeltcheva...... 91

Application of preparative size-exclusion chromatography (SEC) for separation/clean-up of platelets from blood plasma before their chemical investigation V. Bardarov, M. Shishenkov, K. Bardarov...... 97

Preparation of secondary standards for X-ray fluorescence analysis in the production of molybdenum concentrate M. Mladenov, J. Jordanov...... 103

Guide for Authors...... 109 Journal of the University of ChemicalI. Batchkova Technology and Metallurgy, 47, 1, 2012, 3-30

ADDVANCED APPROACHES OF INDUSTRIAL INFORMATICS IN THE CHEMICAL INDUSTRY (REVIEW)

I. Batchkova

University of Chemical Technology and Metallurgy Received 10 December 2011 8 Kl. Ohridski, 1756 Sofia, Bulgaria Accepted 06 February 2012 E-mail: [email protected]

ABSTRACT

The market globalization and the rapid development of ICT force the chemical industry to new challenges that require enterprise reengineering, introducing new forms of organization, as well as quickly and successfully innovation processes. Main means to cope with these new challenges are the approaches, methods and tools of industrial informatics, which are characterized by great diversity and versatility. This paper aims to provide an overview and analysis of some of the most successful achievements of the industrial informatics in the chemical industry sector in order to achieve better selection and development of combined approaches and scenarios enhancing the agility and efficiency in this sector. Keywords: Industrial informatics, chemical industry, architectures, modeling, control, software

INTRODUCTION and innovations in the enterprises of chemical industry should strive to achieve the following goals: The chemical industry is a significant part of • Satisfaction of customer needs of a wide range European economy, producing a variety of intermediate of high quality products at lower prices and in smaller and final products including base, specialty and consumer quantities; chemicals and pharmaceuticals, with over 1.2 million • Reducing the time of the products to reach the employees in European Union. According to the KRMG market, which is related to shortening the time for their analysis of the future of European chemical industry design, manufacture and supply i.e. ensuring a shorter [1], in 2007, 12 of the 30 leading chemical companies life cycle of manufactured products; are headquartered in Europe and they produce 10 % of • Production of goods with lower costs and the sector output. This analysis shows also that the improved quality through continuous implementation European chemical industry is facing new challenges and use of new technologies into existing production that require decisive steps to fast reengineering and systems. restructuring of the chemical plants due to changes The above mentioned goals, briefly summarized imposed by the global competition. For example, the as “producing the right product at the right quantity of study argues that in the coming years the restructuring right place and at the right time” require a new type of or closure of 40 crackers will be required worldwide, production systems, implementing new forms of business 14 of which are located in Europe with 26 % from the organization and new strategies for designing new European production. It is claimed further that 10 of products, their production and control, enabling rapid 17 plants for ethylene glycol are doomed to design of new products, agile plant reconfiguration and ineffectiveness, representing 65 % from the European quickly introduction of new technological methods. One production of this product. The reengineering activities of the most comprehensive reports discussing the future

3 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012 challenges and development trends in chemical indus- der to create adaptive, digital, networked and knowl- try, the “Technology vision 2020” [2], emphasises four edge-based process industries. Unfortunately, despite the main technical disciplines crucial for the development importance of industrial informatics for the practice, of process industry, namely: new chemical science and there are only few publications summarizing and ana- engineering technology; supply chain management; lyzing its application in the chemical industry sector. information systems; manufacturing and operations. Some of the most popular are proposed by Marquardt The rapid development of information and [4], Grossmann [5] and Shah [6]. communication technologies (ICT) and computer The finding noted above initiates the topic on science affects differently and to varying degrees all the which this article is focused, namely to present a review above mentioned disciplines, offering a wide range of discussing the state of the art and future trends in the approaches, methods, tools and devices. The intensive development and application of the approaches, methods use of ICT and computer science in industry leads to and tools of industrial informatics for the field of formation of a new branch of computer science, so- chemical industry in order to support the improvement called “industrial informatics”. Industrial informatics of competitiveness and agility in this industry sector. In denotes the study of design, use and innovation of ICT the next part of the paper a short analysis of the main and computer science in industrial contexts [3]. In recent tasks and topics of industrial informatics and their years, industrial informatics penetrate intensively and importance for the agility, dependability and efficiency multilaterally in chemical industry, filling the gap that of industrial enterprises are presented. Part 3 of the paper exists for decades between products, processes, discusses the main tasks of the disciplines, selected as equipments and business by providing relevant data for the most important for the future development of chemi- real time control and decision making. This is possible cal industry and the contribution rate of industrial thanks to technological advances in areas such as: informatics for their solving. In part 4 different communications and internet technology, intelligent approaches, methods and tools of industrial informatics measurements and digitization processes. Industrial with contribution to the integration processes are informatics and the closely related to it control and analysed and discussed. Part 5 discusses technologies automation systems allow the achievement of a radical which support the engineering processes in chemical transformation of the main production processes in or- industry and Part 6 is connected to the formal ap-

Fig.1. Foundation of Industrial Informatics.

4 I. Batchkova

proaches, applied to increase the correctness of the de- • Technologies supporting integration; veloped software systems. Finally some conclusions on • Technologies supporting engineering. the future trends in the use and application of indus- Although these technologies are developed from trial informatics in the chemical industry are presented too long in the fields of theoretical informatics, their and new research ideas are suggested. industrial application and integration with other techniques and approaches requires special attention and SHORT OVERVIEW OF INDUSTRIAL further research. INFORMATICS The research field of “Industrial Informatics” is MAIN TASKS FOR IMPROVING THE relatively young, but with already established global COMPETITIVENESS OF CHEMICAL INDUSTRY positions and is essential to address challenges in the In the early 21st century, chemical industry is advanced development of economy and society and for facing new challenges that are a consequence of the the development and success in the field of automation increasing globalization of markets, societal demands and control [7]. The tasks, which are solved by industrial for higher environmental performance, financial market informatics, can not be solved by means of automation. demands for increased profitability and capital In this sense, industrial informatics supplements the productivity, higher customer expectations, and changing control and automation by offering new methods and work force requirements [2]. The “Technology vision approaches combining knowledge from different 2020” investigates the following four technical disciplines of computer science and information disciplines which are crucial for the future development technology (Fig. 1). The approaches, methods and tools of chemical industry: of industrial informatics integrates computing, • New chemical science and engineering tech- communication, and storage capabilities with the moni- nology; toring and/or control functions and activities of the el- • Supply chain management; ements in the physical world in order to achieve de- • Information systems; pendable, safety, security and efficient real time sys- • Manufacturing and operations. tems. It requires good knowledge in the areas of theo- As shown in Fig. 2 between the selected disci- retical and technical informatics, and a good founda- plines there are cross-links, which represent the mutual tion in control theory, systems engineering, electrical influence and the complex relationships between them. engineering and mathematics. The approaches of indus- In the next paragraphs we will discuss the main tasks that trial informatics are used at all hierarchical manufac- need to solve each one of these disciplines and the corre- turing levels - from process level through all levels of sponding technologies of industrial informatics, which control to the business level to solve the tasks of con- facilitate in highest degree their successfully solution. trol, planning, processing, etc. Also different methods and tools of industrial informatics are used to solve various tasks at the different production system and product life cycle phases, as for example requirements definition, design, implementation, integration, etc. All this leads to the variety of technologies, methods and tools that are used in this area and creates difficulties connected to their systematization and classification. Therefore, without claim to be exhaustive this paper presents only technologies of industrial informatics which are crucial for the chemical industry and are sub- ject to long-term studies by the author. In line with this, the key enabling technologies constituting the founda- tion of industrial informatics may be summarized in the following groups: Fig. 2. Foundation of Industrial Informatics.

5 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

Main tasks and supporting technology of the new • The technologies of industrial informatics may chemical science and engineering technology help also to improve the systems integration capabilities The Technology Vision – 2020 report defines needed to effectively apply engineering technologies. The three main areas for the chemical science of future: process control and optimization needs to be integrated chemical synthesis, bioprocesses and biotechnology, and from one site to the process development and to the materials technology. The main tasks in these fields plantwide and multisite implementation from another site. include the development of: • Development of instrumentation interfacing • New catalysts and reaction systems based on standards in order to enable the use of distributed ana- economical and environmentally safe processes with lytical networks and more efficient data acquisition and lowest life-cycle costs; control systems. Traditional quality control laborato- • New cost-effective techniques to create a ries will be replaced by real-time, continuous, in-pro- broader variety of molecular architectures in alterna- cess measurement of composition and properties. tive reaction media and for stereo-specificity, or preci- • Open architecture software tools will be devel- sion in spatial arrangements of molecules; oped to permit rapid “plug-in’’ integration of tools and • Chemistry for the use of alternative raw mate- data from a variety of sources, along with a consistent rials; user interface for rapid selection of chemistry; physical • New synthesis tools to efficiently create multi- property databases (thermodynamic, kinetic, and trans- functional materials that can be manufactured with at- port) or ways to compute values for commercially im- tractive economics; portant systems; rapid product and process development • Improved performance of biocatalysts and bio- that optimizes technology options, manufacturability, chemical processing; environmental and safety issues, and financial perfor- • Technology for prediction of materials proper- mance; solids processing; plantwide process control and ties, for precise manipulation of material structures and multisite optimization; and production planning, sched- enhanced performance in materials so as for integrated uling, and optimization of value-adding chain. materials and processes for reuse and disassembly; • Development and use of information infrastruc- • Development of routes for step change improve- tures, which permit data sharing regardless of geographi- ments in performance of materials systems with the use cal location with sufficient safeguards to protect pro- of new additive technology. prietary information. For example, well-maintained na- The most important technologies of industrial tional databases and high-speed networks will make it informatics supporting the successful solution of these easy to share results and avoid duplication. tasks may be summarised as follows: • Improving the design principles, tools, systems, Main tasks and supporting technology for the supply and infrastructures for applying computational technolo- chain management gies. Computational technologies have a broad range of The supply chain links suppliers, producer and applications, from molecular modeling and computa- customer to work together. As shown in Fig. 3 the supply tional fluid dynamics through the process simulation to chain starts at the molecular level with chemicals that operations optimization and control. Increased support must be discovered or synthesized [8]. Over the next steps, is required for experimental validation of computational molecules are aggregated into clusters, particles or films, results. which may be single and multiphase systems becoming • Supporting the development of computing plat- the final form of a macroscopic mixture. Next step in the forms including high-performance desktop workstations, chain is the design and analysis of production units, large fast vector-processor machines or highly parallel followed by integration into process flowsheet. Finally, processors, which require major advances in support tools the designed process becomes part of a site with multiple for system operation and programming, parallel numeri- plants connected by suppliers, warehouses and distribution cal algorithms and template applications; and fully devel- centers, which ultimately determine the commercial oped, optimized, and supported end - user applications. enterprise.

6 I. Batchkova

The main tasks of Supply Chain Management Main tasks and supporting technology for informa- (SCM) are connected with planning and processing orders; tion systems handling, transporting, and storing all materials purchased, The third major discipline concerns information processed, or distributed; and managing inventory. The op- systems, which are strongly influenced by the methods erating environment is characterized by massive amounts and tools of industrial informatics due to the direct of information which must be shared among participants in connections of these systems with the production. The the global supply chain, using systems and data that most Technology Vision considers the information systems often are technically incompatible and inefficient. The in- from three different perspectives: as infrastructure, with formation processing is one of the six drivers of supply chain applicability in business/enterprise management and the efficiency (market globalization; growth of free trade; regu- use of information systems for products and process latory restrictions; transportation; environmental, health, and design and development. safety concerns; and information processing). The main The trends in infrastructure are related to the supporting technologies of industrial informatics are con- use of open systems. There are significant problems in nected with solving the following tasks: the transmission and exchange of information, lack of • Development of information systems (software communications between systems. Crucial in respect to and hardware packages) for SCM based on harmonized this is the improvement of the software and hardware communications, data transmission, and information quality and reliability, resulting in the development of processing to support the global supply chain activities; appropriate interfaces between databases, development • Achieving interoperability in information and use of technologies for data compression, automatic technology to enable connectivity between various computing devices and software systems. hardware manufacturers, software and communications The development of business and enterprise man- products; agement systems, so as information systems for product • Development of large, integrated, computer and process design and development should be based networks for the chemical industry; on standards for infrastructure and data exchange with • Supporting the modeling of safety and environ- a high degree of information integration ensuring mental protection among participants in the global sup- plantwide information system and trends to develop and ply chain. use of Knowledge Based Systems (KBS).

Fig. 3. The “chemical supply” chain [5].

7 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

Main tasks and supporting technology for manufac- internal (Fig. 4-A) and external (Fig. 4-B). The impor- turing and operations tance of these forms is determined by the characteristics Manufacturing operations require advances in the of the industry sectors affected by the integration pro- following areas: relationships with customers; cesses. The external integration is more characteristic for production capability; information and process control; industries consisting mainly of small and medium enter- engineering design and construction; integration of prises such as furniture industry, while the internal inte- suppliers; and global expansion. The main tasks are gration is particularly important for the large companies connected with improvement and reliability of supply and corporations such as automotive, aerospace, ship- and responsiveness to changing customer requirements building, chemical industry, etc. [9]. The range of inte- and development and applying of new technologies for gration processes is presented in general concerning the agility to quickly respond to change, achieve reconfigu- three pillars of supply chain activity - technical equip- rable manufacturing plants, integrated management of ment and hardware used; application systems and R&D, production capability, quickly access to product and and manufacturing and administrative activities. The in- process information, rapid introduction of new prod- tegration process comprises three main tasks: adaptation ucts. Other tasks include shorten the time to design and of existing software, creation of new software for integra- build new plants and information and control systems tion links (as for example middleware or life cycles ap- based on modular components and achievement of glo- proach) and standardization of the integration processes. bal compatibility of information systems. Standardization in industrial informatics TECHNOLOGIES OF INDUSTRIAL INFORMATICS The main factors influencing the increasing de- SUPPORTING INTEGRATION velopment and application of ICT standards in Integration processes are processes of creation integrated systems are [10]: of new distributed systems to improve efficiency of their • The need for sharing of information between work. In distributed systems, each component is regarded the supply chain participants; as a subsystem. The functions of the integrated system • The need to manage this shared information in are a set of functions that are realized by the individual heterogeneous environments within and outside the en- subsystems. There are two basic forms of integration: terprise;

Fig. 4. Internal and external integration.

8 I. Batchkova

• The need for publication and dissemination of A. Standards for exchange and sharing of data information between virtual enterprises and supply In the stage of increasingly evolving integration, chains; the problem for standardization of product data • The need to integrate information from various exchange and sharing becomes critical. The main sources; obstacles in this process are: • The need for rapid preparation of reports and • Firms produces the same product on their own analysis based on shared information, integrated rules; information for long-term forecasting, asset management, • They use various information and control sys- market analysis, manufacturing and other indications; tems requiring a different set of data to achieve the same • The need to integrate information from techni- goal. As a result, the information received by one com- cal and non-technical sources. puter system can not be readed by the other; The systematic nature of advanced technologies, • The product life cycle is much longer than that manifested in their sophistication and complexity, leads of the applyed computer systems and hardware [12]. to an increasing number and variety of standards. They There are several technical approaches for re- in turn have an invaluable impact on the development moving these barriers [9]: of industrial sectors and market structure. Standards are • Use the same software systems; a form of technical infrastructure from major public • Development of translators between systems importance. They are constructions based on consequ- that share data, the total number of required translators ential collective choice, and contract for solving repeti- is N. (N-1) where N is the number of systems (Fig.5a); tive problems. Standardisation is a process of creating a • Development of a neutral format for data ex- balance between user requirements, technological ca- change and sharing through which each system can be pabilities, production costs and regulations imposed by connected. This requires only two translators for the the government to benefit of society [11]. Standards are system (Fig. 5b). pervasive infrastructures influencing in a complicated ISO10303 standard, known as STEP (Standard way the modern economics. In terms of increasing com- for the exchange of Product Model Data) is an plexity and holistic nature of the economic, the struc- international standard for computer-interpretable turing and development of standards is difficult and representation and exchange of product data. Its objec- complex task. tive is to provide a neutral mechanism for describing

Fig. 5.

9 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012 product data throughout the life cycle of a product, inde- • AP231: Process Engineering: Process Design pendent from any particular system. These descriptions and Process Specifications of Major Equipment [17]. enable access, use and exchange of product data through- PIEBASE (Process Industry Executive for out the product life cycle, independent of any particular achieving Business Advantage using Standards for data application or operating system. The standard is organised Exchange) is an international group from the Process as a series of Parts, each published separately. The Parts Industries, working towards the delivery and use of of this International Standard fall into one of the follow- internationally accepted standards for information ing categories: description methods, integrated resources, sharing and exchange [18]. The main goal of PIEBASE Application Protocols, Abstract Test Suites, Implemen- is to deliver a roadmap that identifies the data exchange tation Methods, and Conformance Testing [13]. Datasets standards required by the Process Industries. As part of for certain products are divided into interdependent com- accomplishing this goal, an activity model based on ponents - application protocols [14], with clearly defined IDEF0 is developed providing a common understanding interfaces. They are created by application experts work- of the industry work processes and information ing in the field. Integrated resources have an important requirements. The PIEBASE activity model links each role for information provision and interpretation and are of the above mentioned STEP protocols and their activ- an integral part of standard. ity models together and relates them within a single There are currently three ISO STEP (Standard common framework. for the Exchange of Product Model Data) Application Protocols (APs) under development specifically for the B. Standards for Enterprise integration Process Industries: Standards supporting the Enterprise Integration • AP221: Functional Data and Schematic Rep- (EI) may be classified in three major clusters concern- resentation for Process Plant [15]; ing the following aspects of standardization: enterprise • AP227: Plant Spatial Configuration [16]; models, languages and frameworks for enterprise mod-

Fig. 6. Standards for enterprise integration [19].

10 I. Batchkova

eling and IT services and architectures. Enterprise mod- various programming environments as OAGIS [27], or elling (EM) is perceived as a pre-requisite for EI and execution and integration services as EMEIS of ENV13550 must provide a set of common languages to describe [28]. The cluster of enterprise models standards contrib- various aspects of the enterprise at different abstraction utes to the development of reusable models (or partial levels and from different view points [19, 20]. The models) representing parts of enterprise structures in frameworks for EM, shown in Fig. 6, define concepts terms of processes, information, resources. Some of the and principles in terms of architecture and methodology well known and applied enterprise models are the Refer- for enterprise modelling and engineering. For example ence Model for Shop Floor Production Standards (ISO ISO 14258 (1998) standard defines concepts and rules 10314) [29], the models and interfaces suggested in ANSI/ for enterprise models [21] and ISO 15704 (1998) – de- ISA-dS95 standard for Enterprise-Control System Inte- termines the requirements for Enterprise Reference Ar- gration [30] and also the models of ISO 15531 for Manu- chitecture and Methodologies [22]. The second part of facturing Management Data Exchange (MANDATE) [31]. this cluster covers different enterprise modeling languages, In the next section the scope and models of ANSI/ISA- which may be used in combination with the standards dS95 standard will be presented in more detail. mentioned above. For example the ISO 18629 (2001) standard defines the Process Specification Language (PSL), C. ANSI/ISA-S95 which is formal language [23]. Another formal language ANSI/ISA - S95 standard provides standard is EXPRESS, proposed by ISO 10303-11 (1992) standard, models, terminology and consistent set of concepts for that is used to product data modeling [24]. defining the interfaces between an enterprise’s business The second cluster covers standardization at ar- systems and corresponding manufacturing control chitecture/platform level including different reference systems, as well the activities of manufacturing opera- architectures and models such as TOGAF [25], ODP [26], tions management [30]. The suggested functional enter-

Fig. 7. Functional enterprise - control model according ANSI/ISA-S95.

11 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012 prise-control model is composed of 12 classes of dif- S95 standard detailed description of the suggested object ferent set of functions as shown in Fig. 7. The informa- models and their attributes is given. tion flows between the different functions are defined The activities associated with production con- and categorized in three main areas: information re- trol (manufacturing operations management) defined in quired to producing a product, information about the Part-3 [34] of the standard are consistent with the Part- capability to produce a product and information about 1 data models definitions. Manufacturing operations actual production of the product. Their object models management is subdivided into four categories: are depicted using the Unified Modeling Language production operations management, maintenance (UML) notational methodology [32]. operations management, quality operations management, The production capability information is presented and inventory operations management. A general activity as a collection of information about all resources for model for all these categories is proposed, with the purpose production for selected times including information about to identify all possible data flows within manufacturing equipment, material, personnel, and process segments. It operations. As shown in Fig. 7, eight main activities are describes the names, terms, statuses, and quantities of defined: detailed scheduling, dispatching, resource man- which the manufacturing control system has knowledge agement, tracking, definition management, data collec- and may be seen as a “vocabulary” for capacity scheduling. tion, execution management and analysis. The product definition information is information shared Using the ANSI/ISA - S95 standard different spe- between production rules, bill of material, and bill of cific ontologies supporting the development and integra- resources. The production information is defined in two tion of enterprise information systems at the different models: requests for production (scheduling information) enterprise level may be generated. They will facilitate and responses to the requests (performance information). and standardize information exchange between the in- A production schedule is made up of one or more formation and problem solving systems and in particular production requests and production responses are the will make possible building of more flexible information responses from manufacturing that are associated with a systems for scheduling and planning, which will manage production request. In Part-2 [33] of the ANSI/ISA - successfully rising process and market requirements.

Fig. 8.The philosophy of ANSI/ISA-S88.

12 I. Batchkova

D. Standard for batch control ANSI/ISA-S88 the process view, design starts with a process model In the process industry ANSI/ISA S88 standard containing the (abstract) chemical knowledge of the [35, 36] and its IEC standard equivalent IEC 61512 [37] process to be realized. From general recipe to control provide domain specific models for design and control recipe this model is stepwise substantiated (i.e. adapted of batch production processes. The standard is shortly to the batch plant). The resulting control recipe de- known as S88. Its structure gives a frame for clearly scribes which actions have to be taken in which order definition of processes and product requirements. The to reach the desired process. The control view (or equip- modular structure of the standard is applicable to all ment view) is described with the physical control model, types of control systems and promotes reusability dealing with sensor and actuator signals, and its abstrac- through breaking up the systems into smaller tions, equipment control and control recipe. The con- components. The standard provides a guideline for trol recipe is the meeting point of the two points of designing batch control applications for manual as well view. As all models in S88, the control recipe is built in as automatically controlled processes. The three con- a hierarchical way. Phases are considered as elemen- trol types in batch manufacturing are basic, procedural tary steps in the control recipe and cannot be decom- and coordination. These control types are applied to posed. A phase describes a basic function of the given the control and equipment modules. To examine the plant such as dosing, stirring or heating. Collections of relation between equipment and procedural control, S88 phases performed in a specified order (sequential and/ defines a hierarchy of four types of recipes, i.e. general, or parallel) are operations. Operations are organized in site, master and control recipe. unit procedures which are assigned to a specific class of The main idea of S88 is to separate product knowl- plant units. A unit procedure describes a series of edge from the equipment used (Fig. 8). To describe a operations providing a high level function. Finally, unit batch process in different grades of detail from chemi- procedures are combined to procedures describing the cal and control engineering points of view, S88 pro- interaction of plant units. S88 introduces the semi- poses a set of seven models as shown in Fig. 9. Taking formal design language Procedure Function Chart

Process General Site Master Control Equipment Physical Model Recipe Recipe Recipe Recipe Control Model

Equipment Process Process Procedure Procedure Procedure Procedure Procedure Cell

Equipment Process Process Process Unit Unit Unit Stage Stage Stage Procedure Procedure Unit Procedure

Process Process Process Equipment Equipment Operation Operation Operation Operation Operation Operation Module

Process Process Process Equipment Control Phase Phase Action Action Action Phase Module

Process View Procedural Equipment View Control

Fig. 9. Process and equipment view [37].

13 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

(PFC). It enables the designer to graphically describe CAPE-OPEN components on the base of software agent the organization of operations, unit-procedures and pro- is developed [43]. The developed and applied ontology cedures using basic elements. OntoCAPE defines the chemical engineering and mod- eling concepts required for specifying desired software E. CAPE-OPEN components. Thus OntoCAPE enhances CAPE-OPEN The main goal of CAPE-OPEN initiative is to with semantic information that helps developers when conceptually design and define interfaces for the major working with complex projects. In order to disseminate components of a process simulator in order to achieve the software standards in the CAPE domain, the stan- interoperability and reusability of these components [38, dardization body CO-LaN (CAPE-OPEN Laboratories 39]. The interface specification of components is based Network) is established. The main goal of the CO-LaN on two competing interoperability standards COM [40] is to support the global acceptance of CAPE-OPEN stan- and CORBA [41, 42]. The CAPE-OPEN project is the dards that will give plug and play capabilities of process first effort of its kind in the domain of chemical modelling components into process simulators [44]. engineering. It has been funded by the EU for duration of 30 months starting in January 1997 [39]. To over- Life cycle approach come the bottlenecks in interoperability, reuse, and in- The main purpose of using the life cycle models novation, the CAPE-OPEN project aims at defining a of systems is to deal with complexity and change. The system of standard interfaces. In the frame of COGents model of the system life cycle is a set of activities and project a framework for combining and assembling links between them to support the development of a

Fig. 10. Taxonomy of the system life cycle models.

14 I. Batchkova

system (software, hardware, production). The method- are. To some extent, this problem is solved by evolu- ology for development of a system is a set of methods tionary models [53]. For these models, not all of the for constructing models applicable to the life cycle of requirements to the final system are known at early de- the system. The process of developing a system com- velopment stages. They are identified and subsequently prises several stages, corresponding to one or more tasks improved and extended by successive steps until an ac- associated with system development, such as the defini- ceptable solution. Some of the methods of this type are tion and requirements analysis to system design (rough the spiral process model [54] and the model based on or detailed), coding, integration, testing, delivery, main- components [55]. The main disadvantage of evolution- tenance and training. The life cycle model also defines ary models is that they can cope with the changes be- the principles and guidelines set out in separate phases. tween phases, but changes within a phase are not per- There are many different life cycle models and that con- mitted. siderable diversity is determined by too many modifi- Currently the most promising and used models cations to the well-known basic models namely: sequen- are the so called agile life cycle models that attempt to tial, incremental, prototypes, evolutionary and “agile” reduce risk and complexity of the development process models. Fig. 10 represents an attempt to establish a classifi- by rapidly changing requirements. The basic principles cation of some of the most common life cycle models. for achieving agility, as set out in “Manifesto for agility” The sequential models are based on a series of [56] are: constantly supply of finished product, sequential activities and in their original versions each continuous cooperation with customers, continuously activity must be completed before the next. These mod- updates in accordance with changes, worthy participants els are suitable for projects that are clearly defined and and their interactions, product simplicity, requirements understood, and require experienced developers. They satisfaction. The agile methods are a collective term for are one of the most popular models used so far and a variety of new models in the development of systems. some of their basic and most commonly used agents Review and some comparisons of these models are are: total life cycle model [45], the waterfall model [46] presented in [57] and [58]. Some of the most popular and V-model [47]. Some new anhancements based on “agile” life cycle models, as shown in Fig. 10 are: Adap- iterations and feedbacks are known and very popular tive Software Development (ASD), Feature Driven De- [48, 49]. Models based on prototypes are trying to im- velopment (FDD), Crystal Clear (Crystal), Dynamic prove the sequential models in respect to the require- Software Development Method (DSDM), Scrum, ments by achieving rapid design of a prototype directed Extreme Programming (XP), AMDD, Essential Unified to those aspects of the system that are visible to the Process (EssUP), Consolidated Agile Framework (CAF) customer. The prototype is evaluated by customers and and others. used as a basis for improving the requirements [50]. The key feature of incremental models according [51] Reference frameworks and architectures is that they allow separation of projects in different de- The framework is seen as a logical structure for velopment stages each stage adds new features and is classifying and organizing complex information. The based on sequential life cycle model. There are two main architectural framework is a set of predefined views, types of models: concurrent and with consistent deliv- interests, key stakeholders and rules of conformity of ery of functionality. The model based on Rapid Appli- views that are set to capture the usual description of the cation Development (RAD model) [52] illustrates the architecture in certain areas or user communities [59]. potential for creation of various models by modifica- Some of the most popular frameworks within compa- tion and combination of basic types. It is based on a nies are: CIMOSA, GRAI-GIM, PERA, Zachman, sequential approach by also including elements of in- ARIS, C4ISR/DODAF and the generalized framework cremental methods and methods based on prototypes GERAM. One of the most popular frameworks is “The with the goal to extremely reduce the development time. Open Group Architecture Framework” - TOGAF. In The main disadvantage of all of the above men- 1987 John Zahman [60] proposes a framework of the tioned methods is that they do not fully meet the evolu- enterprise architecture, which has the form of a matrix, tionary nature of systems, regardless of what kind they representing the logical scheme for classifying and or-

15 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012 ganizing the descriptive representations of an enterprise tem consisting of components, relationships between that are relevant to management of the enterprise and them and to the environment and also the principles the development of information and control systems. underlying their design and development. The view is The elements of the framework can be viewed as an defined as a description of a system from the perspec- organizational scheme for all types of meta-data included tive of a variety of related concepts. As such, the view in the construction and use of information and control consists of one or more models. systems. Zachman framework is based on the fact that the description of things depends on the perspective Interoperability and knowledge-based systems (viewpoint) from which they are treated and the ques- The knowledge systems based on ontology’s play tion we ask ourselves when we want to describe them. a very important role in achieving enterprise integration. Thus Zachman consider two-dimensional matrix for Ontology’s make explicit assumptions in the domain of classifying architectures. The rows of this matrix repre- subjects and support the development of sharable com- sent different levels of abstraction in system develop- mon information structures from people or intelligent ment. They classify the perspectives of the enterprise in agents. Their usage allows splitting the operational scope, business (owners), system (designers), technol- knowledge from the domain knowledge and opens the ogy (constructors), detailed representation (traders) and possibilities for domain knowledge analysis and reuse. functioning enterprise. Columns reflect the different That is why in the field of enterprise management there perspectives involved in the phases of development clas- are a lot of investigation in development and applica- sified as answers to the questions “Who,” “How,” “What” tion of different kinds of ontology’s. The most popular “Where,” “When” and “Why.” The set of 36 elements and used one can be classified in two main groups: do- of the matrix uniquely classify the individual parts of main (enterprise) ontology’s and meta-ontology’s, as the enterprise. The author presents the framework as a shown in Fig.11. classification scheme rather than a methodology for The ontology’s from the first group are used as a implementation. In 1992, Zachman and Sowa expand basis for the development of different reference archi- the framework and represent it as architecture of an tectures and models for integration in the enterprise information system [61]. engineering. The most popular are: the Enterprise The IEEE 1471 standard from 2000 [62] is re- Ontology ÅO [63], TOVE [64] and REA [65]. They use lated to the recommended practice for architectural different methodologies, levels of formalization, lan- description of software intensive systems and defines guages etc. EO and REA are semi-formal, based on architecture as the fundamental organization of a sys- methodology from Uschold and King [66] combined

Fig. 11. Taxonomy of existing Enterprise ontology’s.

16 I. Batchkova

with this of Gómez-Pérez [67]. As semi-formal they integration in different applications. PSL represents pro- use natural language for its concept dictionaries and a duction processes and PIF – business processes. At the small number of axioms in Ontoligua language. TOVE moment these languages are united. is fully formal and represents an integrated model speci- Essential stage in Enterprise integration is the fied in KIF and implemented in Prolog, and is based on task of optimal planning and integration between the the methodology of Grüninger [68]. In TOVE concepts planning system and other business and management are grouped in taxonomies, and represented by vari- systems. The planning process includes varieties of tasks ables and constants, while attributes and relations are for planning of resources, materials, processes and tasks represented with first-order predicates. EO and TOVE for different periods of time, management of execution are combination from different ontology’s. For example of these plans and their adaptation to inner and outer EO has 5 basic classes, which represent different enter- changes. The planning tasks can be categorized as: prise aspects: Meta ontology, Activities and processes, strategic, tactic or operational, depending from given Organization, Strategy and Marketing. REA uses a dif- goals and execution time. Connections between strate- ferent approach. It is based on theoretical model from gic, tactical and operational levels can be facilitated using the field of accounting and economics theory with do- decision making systems, which use ontology’s. Research main extensions. and developments in this field can be grouped in 2 main The representatives of the second group such as groups: Ontology’s for planning and ontology’s for IDEFF5 [69], PSL [70], PIF [71], etc. are meta- scheduling. ontology’s and they assist in building languages for de- Planning ontology’s are mainly domain ontology’s veloping and maintaining of ontologies in the domain based on standard meta-ontology’s and languages (PIF, “enterprise”. IDEF5 uses 2 languages – graphical PSL, KIF, OWL) like SPAR, DDPO, TOVE, TMDA notation and KIF based language, which differs from and PLANET, and ontology’s used for developing of KIF and Ontoligua in using meta-concepts “process” shared models for planning and plan representation like and “state” which make IDEF5 applicable in develop- PSM, CPR, ACT, OPT developed under various ing enterprise ontology’s. PSL and PIF are KIF based DARPA/ARPI projects [72-74]. The second group of languages that specify enterprise processes, for easier ontology’s uses specially developed planning languages

Fig. 12. Taxonomy of existing ontology’s for planning and scheduling.

17 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

(STRIPS, CPT, ACT, PDDL, KRSL, TF), which do not provides an architectural approach that brings the flex- allow integration of different ontology’s with given one ibility and agility required by today’s global business as in the taxonomy, shown in Fig.12. The scheduling environment. SOA addresses the business demand for ontology’s are JATO (Job assignment task ontology) [75], applications to be responsive to business needs and to MULTIS [76], OZONE [77] and TMDA [78, 79]. adapt to dynamic business environments. Different ar- The main shortcomings of these ontology’s are the chitectures and models for Service Oriented Architec- insufficient level of formalization and the lack of com- tures are known. Some of the most popular are for ex- plete definitions of the main concepts in the task domain. ample W3C Web Service Architecture, OASIS SOA This lead toward elimination of tasks connected with op- Reference model, ebXML, Semantic Web Services, JINI, timization and deviation from the defined requirements. OSGi etc [83]. All these initiatives are collections of The main conclusion, which may be done from the above best practice principles and patterns in service-oriented summarized state of the art for the development of enter- design. prise ontology’s, including these for planning and sched- uling, is that currently there is not a standardized method- TECHNOLOGIES OF INDUSTRIAL ology for their development. The existing ontology’s cover INFORMATICS SUPPORTING ENGINEERING only particular aspects of subject domains and do not fully represent both domain and domain tasks. Information modelling The concepts as conceptual model, reference Web service and Service Oriented Architectures (SOA) model, ontology, meta-model, meta-meta model are The powerful services in chemical industry are becoming more common in the field of system very important requirement in achieving agility. Different engineering and control system development and approaches are known, which try to solve these tasks. analysis, including all hierarchical levels and The breakthrough directions are the technologies based applications. For this purpose it is necessary to clarify on web services enabling peer-to-peer communications the nature and the relationship between these concepts, between agents. Web services are a response to the increasingly occurring in the field of information and request for creation of distributed computing application control systems. There are different definitions of the environment that is not so difficult to deploy as CORBA, term “model”. Most often, the model is treated as a DCOM or JRMI and would offer greater interopera- representation, description and specification of entities bility. The main aim of CORBA, JRMI and DCOM is or reality with particular goal. One of the most com- to provide a distributed computing environment across prehensive definitions is given in [84] where the model heterogeneous environments, but unfortunately neither is seen as “external and explicit representation of a part of them supports communications at the peer-to-peer of reality as perceived by people who intend to use the scale. The key enablers of the Web services technology model in order to understand, change and control this are XML, SOAP [80], WSDL (Web Services Descrip- part of reality”. For purposes of defining the concepts tion Language) [81] and UDDI (Universal Description, of “meta-model” and “meta-meta-model” the model M0 Discovery, and Integration) [82], and may be created (as shown in Fig. 13) is treated as an abstraction of a from existing systems with a small amount of custom part of reality for certain purposes and is given by mod- code, thereby saving valuable time and money. The de- eling language L0. Models and modeling processes are fined web services use WSDL. The service provider de- based on two basic concepts: conceptualization and ab- scribes its service using WSDL, which is published to a straction, which are immaterial entities that exist only directory of services UDDI. The service determines how in human minds or in the people community. To be to communicate with that service and uses the WSDL documented, disseminated and analyzed, they must be to send a request to the service provider. The service named with specific terms, which mean that a language provider provides the expected response to the service for their representation in a complete, concise and un- consumer. The Service Oriented Architecture (SOA) is ambiguous manner is needed. The relationship between a software model in which the concept of a “service” is language, conceptualization and part of reality that is an abstraction of a function used by an application and ignored by the conceptualization is presented by the so

18 I. Batchkova

known “Triangle of Ullman” [85]. In respect to this, the Like the term “model”, the “meta-model” has conceptual modeling includes the processes of identifi- also been defined in various ways. Some authors con- cation, analysis and description of the basic concepts sider the “meta-model” as a conceptual model of the and constraints in a specific domain using a modeling technique (method) for modeling, others define it by language based on a small set of basic meta-concepts. the term reference model [86] but most simply may be This way the conceptual model is a recommended model, defined as a “model of models” expressed in a particu- representing set of entities (artifacts) by a set of con- lar language [87]. Meta-model is still regarded as a defi- cepts, relationships between them and constraints un- nition of abstract syntax [88]. The so-called “FRISCO der the assumption of “closed environment” (the “closed report” [89] gives a more detailed description that the environment” is a concept of formal logic that defines meta-model is a model of the conceptual basis of lan- the presumption that what is not currently known to be guage, consisting of a set of basic concepts and a set of true is a false). rules defining the set of possible models created with Unlike the conceptual model, the above discussed this language. But a very simple definition may be also concept of ontology refers to the capture of significant that the meta-model M1 is a conceptual model of the entities in the domain, using language for specifying modeling language L0. The modeling language is a well ontology’s, based on a small set of basic, independent understood language (not necessarily formal), which of the domain, ontological categories (forming ontology describes concepts and their relationships for a part of of higher level). Thereby ontology is a shared, descrip- reality. The meta-meta model is a model of the lan- tive structural model representing the reality through a guage for meta-modeling, which can be used to express set of concepts, relationships between them and the con- meta-models. The model of modeling language for meta- straints under the assumption of an “open environment” meta-models is expressed by itself. This way the meta- (“open environment” is the opposite to “closed envi- meta-model M2 is a conceptual model of language for ronment”). meta-modeling L1. The relationships between the so-

Fig. 13. Semantic relation between different types of models.

19 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012 defined model types are presented in Fig. 13, which show advanced approaches for information modeling such as that for a model M to represent an abstraction A, the the object-oriented. modeling primitives of the language L, used to constitute M, should accurately represent the conceptualization of Object-oriented approaches the domain, used to express the presented abstraction A. UML [32] is a general purpose modelling Important is to distinguish between meta- and language, which gives uniform notations and a meta- meta-meta models on one hand and the so known model for object-oriented modelling and software design. reference models on the other hand (Fig. 13). The refer- It does not specify a methodology for software design ence models are general conceptual models that formal- but supports the Model Driven Architecture (MDA) ize recommended practices in the domain. The main task consisting in transformation of different platform of the reference models is to accelerate (rationalize) the independent models towards executable applications. design of different types of models through a common UML aims to provide an integrated modelling framework, solution. The use of reference models are motivated by covering structural, functional and behaviour the paradigm of “design by reuse”. Reference models descriptions. accelerate the modeling process by creating of repositories The main challenges towards a real-time UML of models for important processes and structures. The are connected with the creation of different mechanisms reference model may be defined as a universal conceptual to handle real-time features such as: models of physical model that uses the same semantic level. time, timing specifications, timing facilities, modelling and management of physical resources and concurrency. Structural approaches In the last years there are many working groups The modeling of enterprises has its roots in system developing and using different approaches and modeling and in particular the modeling of information techniques in order to extend the object-oriented software systems. The first methods that were used for modeling development and especially UML to the field of real enterprises are ER approach of Peter Chen [90] and SADT time system modelling and analysis. (Structured Analysis and Design Technology) approach One of the most important profiles of UML, sup- of Douglas Ross [91]. These methods are focused on the porting the modelling of chemical systems is the SysML information and functional modeling, respectively. They [103]. It is a general-purpose modelling language for are the foundation of later emerging structural software system engineering that reuses a subset of the last engineering methods (SSAD, structural methods of UML2.1 version and provides additional extensions Yourdon [92], structural analysis of De Marco [93], through stereotypes, diagram extensions and model li- structured programming Jackson, structural design of brary in order to model a wide range of system engi- Constantine, etc.). A detailed overview of these methods neering problems as for example specifying require- was made by Martin and Makklar [94]. In the early 80s ments, structure, behaviour, allocations and constraints appeared a number of modeling techniques generated by on system properties to support engineering analysis. the idea of creating a computer-integrated manufactur- The reusable subset of UML, known as UML4SysML ing. Some of the most popular are the methodology IDEF includes Interactions, State machines, Use Cases and [95] and methods GRAI and GRAI-GIM [96], based on Profiles. In Fig. 14 the set of SysML diagrams in re- activity modeling. In the 90s, new competitive methods spect to their modelling aspects is summarized. appeared, part of them are process oriented as for example The system structure design is supported by four ARIS [97], CIMOSA [98], IEM [99]) and the other - types of diagrams: Block Definition Diagram (BDD), object oriented methods as OOA [100], OMT [101], etc. Internal Block Diagram (IBD) reinforced by Paramet- until the emergence of UML [32] and MDE-MDA [102]. ric and Packages diagrams. The BDD is an extended The existing modeling languages can be classi- and restricted UML class diagram that is composed of fied in different ways. One of them includes the follow- blocks and relationships between them, such as asso- ing groups: static-oriented, business-oriented, state-ori- ciations, generalizations and dependences. The SysML ented and communication-oriented languages. In this blocks are based on UML classes with extended com- paper, particular attention will be given to the most posite structure and are defined in terms of properties,

20 I. Batchkova

operations and relationships. The IBD captures the in- chine diagram, and use case diagram. The activity dia- ternal structure of a block in terms of properties and grams are used to describe the flow of control and flow of connectors between properties. Three general catego- inputs and outputs among actions. The state machines de- ries of properties are recognized: values, parts and ref- scribe the constructs used to specify state based run time erences to other blocks. Ports are a special kind of parts, behaviour in terms of system states and their transitions. which give access to internal structures from the out- The sequence diagrams describe the communications be- side of a composite object. In SysML, unlike UML, ports tween different structural elements over the time and iden- can also be complex hierarchical structures. There are tify the required relationships and messages. The use case two basic types of ports – standard and flow ports. Stan- diagrams model the relationships between users and a sys- dard ports are the same as in UML2.x and allow speci- tem enabling the specification of requirements to the sys- fying services to or from the environment through the tem behaviour at the different decomposition levels. definition of provided and required interfaces. Flow The Requirements diagrams, which can be presented ports are an extension of UML2.x ports and specify what in graphical, tabular or tree structure format, are used to “can” flow in or out of a block. Different items such as specify different constructs for system requirements and data, material or energy may flow through these inter- to cover the relationships between them. In SysML two action points in order to achieve asynchronous, broad- kinds of requirements are used – functional and perfor- cast, or send and forget interactions. The concept of mance, as they specify the capabilities or the conditions item flows specifies what “does” flow between blocks which must be performed or satisfied by the system. and/or parts and across associations or connectors in a Other modelling capabilities of SysML, not shown particular usage context. Parametric diagrams allow in Fig. 14 are the cross-cutting constructs, such as allo- capturing system properties and constraints in form of cations for connecting the different views, and Profiles & different equations (algebraic, logical, differential, etc.) Model libraries allowing further to customize and ex- and support the analysis and verification processes. Pack- tend SysML to specific applications. SysML includes ex- ages diagrams are the same as in UML2.x and facilitate tensions supporting the causal analysis, the verification the decomposition and encapsulation of the projects. and testing processes and the decision tree development. The Behaviour diagrams incorporate four dia- SysML is developed as an open source specifica- grams: activity diagram, sequence diagram, state ma- tion and the last draft SysML1.2 is submitted to the

Fig. 14. SysML Diagrams [103].

21 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

OMG for technology adoption in June 2010. There are ports. In order to present the dynamical behaviour of numerous modelling tool vendors who have already the plant, the elements of “TankSubSystem” are de- updated their UML2.x based tools to comply with the scribed with their transfer functions as operations using OMG SysML specification as for example ARTiSAN Parametric diagram. Parametric constraints are added to Studio, Rhapsody, MagicDraw, Enterprise Architect etc. the model in order to support the analysis tasks and to [104]. For the case study, described in this chapter Rhap- describe the physical systems with equations. In Fig. 15- sody of Telelogic is used [105]. 5 the parametric constraints for the “TanksSubSystem” The advantages of the UML profile for system are presented in Parametric Diagram through the trans- engineering SysML in comparison to these based on fer functions for tank, input and output valves. The transfer UML2.x may be summarized as follow: functions describing the tank include the following pa- • SysML supports the whole life cycle by the rameters – T, Kin and Kout which are presented as development of control engineering applications from attributes. These parameters may be calculated using the requirements definition to the software indicated equations. implementation; • Based on extended activity diagrams, parametric TECHNOLOGIES OF INDUSTRIAL INFORMATICS diagrams and flow ports and items, the proposed FOR IMPROVING OF CORRECTNESS approach may be applied for continuous and hybrid systems; The information and control systems in chemical • The possibilities to model the physical systems industry are complex reactive real time systems, which in detail enhance the procedures for analysis, testing exhibit various behavioural aspects such as and validation of designed closed loop behavior of the communications between components, state system; transformation inside components, and real-time • Through encapsulated objects in UML 2.x and constraints on the communications and state changes. hierarchical structure, which they gain using SysML Their correctness depends not only on the results it pro- profile the object-oriented systems get near to agent- duces, but also on the time at which the results are avail- based systems; able. There is a need to provide means to specify tem- • In order to improve the modelling processes, poral constraints or time dependent behaviour of the the UML/SysML notations may be mapped to concepts system and its components, and allow properties of safety from IEC-61499 standard. and liveness to be expressed and checked or proved One of the big advantages of SysML profile is through the system development life cycle. A main ap- the possibilities to model physical systems. We will show proach for fulfilling these requirements and making the a simple case study for modelling of physical system systems more reliable is the use of formal approaches consisting of a cylindrical tank, filled with water until at the early life cycle phases of their development. specified level that is registered with a level sensor and is The formal approaches cover two aspects of the de- controlled by changing the input valve position (Fig. 15). signed software systems: formal specification or modelling The static composite structure of the “TankSubSystem” is and formal verification. The formal specification is the first presented with BDD (Fig. 15-2) and shows the physical step in the formal development of a software system and structure in terms of installed equipment. In Fig. 15-4 the expresses a collection of properties that the designed system IBD of “TankSubSystem” is presented, which is com- should satisfy in some formal language, which syntax can be posed by different types of equipments such as input mechanically checked for correctness and at some level of valve, output valve, tank and sensor. Each part is de- abstraction. The properties are most often expressed in some fined uniquely through its attributes and operations. The logic and must have a well defined semantic. Bowen and attributes characterizing “Valve_in” and “Valve_out” are Hinchey in [107] analyse the last trends in the development respectively input/output pressure (Pin/Pout), input/ and industrial applications of formal methods. Some sur- output flowrate (Qin/Qout), Cin/Cout for density of liq- veys on the formal methods for real time systems are pre- uid and Sin/Sout for coefficient of input/output valve. sented in chronologically order by Ostroff [108], Clarke at The connections between parts are modelled as flow al. [109], NASA [110] and Wang [111].

22 I. Batchkova

Fig. 15. UML based modelling of tank subsystem.

Specification techniques can be classified in many The model-oriented languages can be further clas- different ways. Fig. 16 summarizes some of the most sified into state-based and event-based methods. State- promised techniques, where the formal languages are based languages uses explicit states in the model whereas classified in two basic classes: property- and model – event-based describe transitions or events without ref- oriented. All formal methods, which realize formalisms, erence to an explicit state. Examples of event-based lan- suitable for specification of real time systems are shown guages are process algebra, traces and grammars [112]. in grey. Property-oriented languages describe a system The frequently used methods of process algebra such as implicitly through its properties, while the model-ori- ACP (Algebraic Communicating Processes) [113] based ented languages construct an explicit model of the sys- on algebraic semantic, CCS (Calculus of Communicat- tem that is more understandable. Property-oriented lan- ing Systems) [114] based on denotational semantic, CSP guages specify the properties of the system in form of (Communicating Sequential Processes) [115] based on equations or first or higher order logical formulas. The operational semantic, and their enhancement with data mostly used languages are algebraic languages such as types - mCRL (LOTOS) [116] have been extended in a Clear and OBJ, axiomatic languages and different type number of ways to deal with the time and are very ap- of logics [112]. Logics, such as prepositional, first or propriate to modelling highly distributed and concur- higher order logics allow unambiguous specifications rent real time systems [117]. but in order to express timeliness properties they must The State-oriented languages are presented through be extended with some concepts of time. The most widely a wide range of different languages and their modifica- used timed extensions are modal and temporal logic, tions, as for example: the widely used automata (finite, duration calculus, and real time logic [108, 110]. timed [118, 119], I/O automata [120], hybrid [121]), set

23 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012 theory based languages as Z and VDM [122], graphs etc. Fig. 17 taxonomy of the most frequently used methods Different kinds of Petri Nets [123] and the Generalized for formal verification is shown. The methods are di- Nets [106] provide a very flexible and expressive nota- vided in two basic classes – manual/semi-automatic and tion, based on simple formal concepts. The most sig- automatic methods. nificant problem of Petri nets is related to complexity The manual or semiautomatic verification needs expressiveness. There are different Petri net extensions a verification expert with a solid background in logic to specifying real time systems. The time in Petri Nets and formal reasoning and are suitable for analysis of is introduced in different ways: in the Timed Petri Nets small-scale projects. Theorem Proving, the main repre- (TTPN) and Time Petri nets (TPN) the time as delay sentative of this class, uses a logical induction to prove constant or interval respectively is associated to the tran- system requirements, which have been translated into sitions, in the Timed Places Petri Nets (TPPN) as delay complex mathematical equations. For the verification – to the places, in the Time Environment/Relation Petri of real time systems preferable is the second class of Nets (TimeE/R-PN) as time stamp – to the tokens and methods and the selected method depends on the cho- in the Timed Arcs Petri Nets (TAPN) and SIPN as sen specification method and on the importance of the travelling delay - to the arcs [124]. properties, which must to be proved or verified. The verification technique provides means to The logics of automatic verification techniques prove if the formal model satisfies the specification or usually have less expressive power than the logics of formalized system requirements, i.e. if the model is build manual verification techniques. The different Model right. The formal verification approaches are hardly stud- checkers verify properties described using the Tempo- ied in the field of software engineering and their adap- ral Logic formalism. The two mostly applied kinds of tation for the domain of real time systems is needed. In Temporal Logic are Linear Temporal Logic (LTL) [125]

Fig.16. Formal specification methods.

24 I. Batchkova

Fig. 17. Formal verification methods. and Computational Tree Logic (CTL) [126], where the time systems. Instead, there are various specialized tech- time is described in a linear and a branching fashion re- niques appropriate to describe the various aspects of spectively. Model checking can be considered as exhaus- system behavior. tive simulation and its fundamental drawback is the state • The temporal domain strongly influences the explosion problem. In order to overcome this problem expressiveness and usefulness of the formal techniques. several techniques like symbolic model checking, partial A trend is to handle non-behavioural aspects of a sys- order reduction, symmetry reduction and bounded model tem like its performance, real-time constraints, secu- checking have been developed and applied [110, 111]. rity policies, and architectural design. The main conclusions, which may be done from the analysis of the classified methods, are: CONCLUSIONS • There is a plenty amount of formal approaches, methods, languages and tools covering different aspects The review presented in this paper covers the and domains in the software system development. A wide, varied and challenging area for application of the main feature of this diversity is that most of the tech- increasingly more attractive and effective approaches, niques have been evolved independently of each other. methods, tools and technologies of industrial informatics, The current trend today is to integrate different specifi- such as the chemical industry. Outside the study are cation languages each able to handle a different aspect very interesting standards for distributed control and a of a system with the aim to improve and extend their variety of intelligent approaches that combine the power applicability. of Industrial Informatics with that of artificial • Most of the available formal specification and intelligence and cognitive science. So I firmly believe verification techniques have been designed and applied that the topic needs a sequel. to the design phase of software development life cycle. There is a need to extend the research and development REFERENCES of formal techniques supporting the whole software de- velopment life cycle and especially for formalizing the 1. KPMG International, The future of the European system and components requirements. Chemical Industry, 2010, available online at: http:// • For now there is no universal method for for- www.kpmg mal specification able to describe all aspects of real 2. American Chemical Society, American Institute of

25 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

Chemical Engineers, Chemical Manufacturers Asso- mittee Draft ISO TC184/SC4/WG3 N600, 1997. ciation, Council for Chemical Research & Synthetic 16. ISO 10303, Part 227, Plant Spatial Configuration, Organic Chemical Manufacturers Association. Tech- ISO Standard 10303-227, 2000. nology vision 2020: The US chemical industry, 1996, 17.ISO 10303, Part 231, Process Engineering Data: Pro- available online at (accessed 10 February 2012): http:/ cess Design and Process Specifications of Major /www.chemicalvision2020.org/pdfs/chem_vision.pdf Equipment, Committee Draft ISO TC184/SC4/WG3 3. J. Holmström, M. Wiberg, A. Lund, Industrial Informatics: N740, 1998. what we know and what we do not know, in the book: 18. PIEBASE (Process Industry Executive for achieving Industrial Informatics Design, Use and Innovation: Per- Business Advantage using Standards for data Ex- spectives and Services, IGI Global, ISBN13: 97816152 change) Working Group 5 with contributions from 06926, June, 2010. America, Europe and Japan. Proceedings, Version 4. R. Schneider, W. Marquardt, Information technol- 1.0, Issued October 1, 1997, 1-21 pp, 1997. ogy support in the chemical process design life cycle, 19. D. Chen, F. Vernadat, Standardisation on enterprise Chemical Engineering Science 57, 2002, 1763 – 1792. modelling and integration: achievements, on-going 5. I. E. Grossmann, Challenges in the new millennium: works and future perspectives, in Proc. of IFAC product discovery and design, enterprise and supply 10th Symposium on Information Control in Manu- chain optimization, global life cycle assessment, Com- facturing, INCOM’01 (Invited session), Vienna, pp. puters and Chemical Engineering, 29, 2005, 29-39. September 20-22, 2001. 6. N. Shah, Process industry supply chains: Advances 20. F. B. Vernadat, Enterprise Modelling Languages, and challenges, Computers and Chemical Engineer- ICEIMT’97 Enterprise Integration - International ing, 29, 2005, 1225–1235. Consensus, 1997. 7. G. A. Fodor, Formal and architectural competitive fac- 21. ISO 14258, Industrial Automation Systems – Con- tors in industrial informatics, Keynote talk of the IEEE cepts and Rules for Enterprise Models, ISO TC184/ – Industrial Informatics Conference, Berlin, 2004. SC5/WG1, April-14 version, 1999. 8. W. Marquardt, L. Wedel, B. Bayer. Perspectives on lifecycle 22. ISO 15704, Requirements for Enterprise Reference process modeling, In: M. F. Malone, J. A. Trainham, & B. Architecture and Methodologies, ISO TC184/SC5/ Carnahan (Eds.), Proceedings of the AIChE Symposium WG1, N423, 1998. Series, No. 323, Vol. 96, 2000, 192–214. 23. ISO 18629-1 CD, Industrial automation systems and 9. Funstep Organization, http://www.funstep.org/funstep/ integration, Process Specification Language (PSL), ecos/ecos.htm. Part 1: Overview and Basic Principles, JW8/ISO 10. J. Fowler, SC4 Data Architecture - Integration and 184/SC4/SC5, 2001. Mapping Methodology - Version 1, ISO TC184/ 24. ISO 10303-11 DIS, The EXPRESS Language Ref- SC4/WG10/N55, 2000. erence Manual, ISO TC184/SC4/WG5, N35, 1992. 11. G. Tassey, Standardization in Technology-based 25. Open Group, TOGAF: The Open Group Architec- Markets, 1999, http://www.mel.nist.gov ture Framework, Document No. 1910, Version 6, 12. M. West, Integration of Industrial Data for Exchange 2000. Access and Sharing: Architecture Overview and 26. ISO/IEC 10746-3, Information Technology – Open Description, ISO TC184/SC4/WG10/N254, 2000. Distributed Processing – Reference Model – Ar- 13. ISO 10303, Part 1: Overview and fundamental prin- chitecture, ITU-T Recommendation X.903, 1994. ciples, ISO TC184 SC4, 1994. 27. OAG, OAGIS: Open Applications Group Integra- 14. UKCEB, Step - A key Tool in the Global Market, tion Specification, Open Application Group, Incor- 2000, http://www.graco.unb.br/alvares/DOUTORADO porated, Release 7.2.1, Doc. No. 20011031, 2001. /disciplinas/estudo_dirigido/capp/www.ukceb.org/ 28. ENV 13550, Enterprise model execution and inte- step/pages/stpgolb1.htm gration services (EMEIS), CEN, Brussels, 1995. 15. ISO 10303, Part 221, Functional Data and their 29. ISO 10314 TR, Reference model for shop floor pro- Schematic Representation for Process Plant, Com- duction standards, Part 1 - Reference model for

26 I. Batchkova

standardisation, methodology for identification of sium on Computer-Aided Process Engineering, requirements, 1991. pp. 1021-1026, (2004). 30. ANSI/ISA-S95, Enterprise-Control-System Integra- 44. M. Pons, B. Braunschweig, K. Irons, J. Köller, A. tion, ISA DS95.01-1999, Instrument Society of Kuckelberg, P. Roux, CAPE-OPEN (CO) standards: America, 1999. implementation and maintenance, 2nd IEEE Con- 31. ISO 15531-1, DIS, Industrial automation systems ference: Standardization and Innovation in Infor- and integration - Manufacturing management data mation Technology, 2001, 335-338. exchange - Part 1: Overview and fundamental prin- 45. P.K. Ragunath, S. Velmourougan, P. Davachelvan, ciples, ISO TC184/SC4/WG8, N138 R3.1, 2000. S. Kayalvizhi, R. Ravimohan, A new model (SDLC 32. OMG-UML, OMG Unified Modelling Language Model-2010) for software development life cycle Specification, Version 1.4, 2001, available at: http:/ (SDLC), IJCSNS International Journal of Computer /www.oml.org Science and Network Security, Vol.10, No.1, Janu- 33. ANSI/ISA-S95.00.02-2004-draft, Enterprise-Con- ary, 2010, 112-119. trol System Integration Part 2: Object models and 46. W.W. Royce, Managing the development of large attributes, Copyright by ISA, ISA, Research Tri- software systems, in: Proceedings of 9th IEEE In- angle Park, NC, 2004. ternational Conference on Software Engineering, 34. ANSI/ISA-S95.00.03-2004-draft, Enterprise-Con- 1987, 328-338. trol System Integration Part 3: Activity Models of 47. DIN, Entwicklungsstandard für IT Systeme des manufacturing Operations Management, Copyright Bundes, Vorgehns-model, Teil 1: Regelungsteil, http:/ by ISA, ISA, Research Triangle Park, NC, 2004. /www.V-modell.iabg.de 35. ISA, Batch Control Part 1: Model and terminology, 48. H. Benington, Production of large computer pro- The International Society for Measurement and Con- grams, in: Proceedings ONR Symposium 1956, re- trol, ISA Press, ISA – S88.01-1995. printed in Annals of the History of Computing 5(4), 36. ISA, Batch Control Part 2: The International Soci- 1983, 350-361. ety for Measurement and Control, ISA Press, ISA 49. W. Hosier, Pitfalls and safeguards in real-time digi- – S88.02-1995. tal systems with emphasis on programming. IRE 37. IEC, Batch control, Part 1: Model and terminology, Transactions on Engineering Management, 1961, The International Electrotechnical Commission 99-115. (IEC), Final draft, 1997. 50. V. Gordon, J. Bieman, Rapid prototyping: lessons 38. M. Jarke, J. Köller, W. Marquardt, L. von Wedel, B. learned, IEEE Software, 12, 1, 1994, 85-95. Braunschweig, CAPE-OPEN: experiences from a 51. V.R. Basili, A.J. Turner, Iterative Enhancement: A standardization effort in chemical industries. Practical Technique for Software Development, 39. J.-P. Belaud, B. Braunschweig, Open software ar- IEEE Transactions on Software Engineering, SE-1- chitecture for process simulation: the current status 4, Dec., 1975. of CAPE-OPEN standard, ESCAPE-12, La Haye, 52. J. Martin, Rapid Application Development, 2002. MacMillan, 1991. 40. COM, Microsoft COM Web Site: http://www. 53. H. Vliet, Software engineering: Principles and Prac- microsoft.com/com/default.asp, 1999. tice, Wiley, 2007. 41. CORBA, OMG Web Site: http://www.omg.org/, 1999. 54. B.W. Boehm, A spiral model of software develop- 42. S. Vinoski, CORBA: Integrating diverse applications ment and enhancement, IEEE Computer, 21, 5, 1988, within distributed heterogeneous environments. 61-72. IEEE Communication Magazine, Vol 14, No. 2. 55. A. Davis, P. Sitaram, A concurrent process model 43. B. Braunschweig, E. Fraga, Z. Guessoum, W. for software development, Software Engineering Marquardt, O. Nadjemi, D. Paen, D. Pinõl, P. Roux, Notes, ACM Press, Vol. 19, No. 2, pp.38-51, Apr. S. Sama, M. Serra, I. Stalker, A. Yang, CAPE Web 1994 Services: The COGents way, In European Sympo- 56. Agile, Manifesto for Agile Software Development,

27 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

http://agilemanifesto.org, Last access on March, 2010. Task taxonomies for knowledge content D07, De- 57. P. Abrahamsson, O. Salo, J. Ronkainen, J. Warsta, liverable to the METOKIS - ICT 507164 project of Agile Software Development Methods, Technical the 6FP, 2004. report, VTT Publications 478, VTT, Finland, 2002. 73. A. Tate, Towards a Plan Ontology, AI*IA Notizie 58. B. Boehm, R. Turner, Balancing Agility and Disci- (Journal of the Italian Association for AI), 9, 1, pline. Addison-Wesley, 2003. 1996, 19-26. 59. H. Mili, M. Fayad, D. Brugali, D. Hamu, D. Dori, 74. Rajpathak D., Motta E., An Ontological Formaliza- Enterprise frameworks: issues and research direc- tion of the Planning Task, International Conference tions, Software - Practice and Experience, Softw. on Formal Ontologies in Information Systems, Pract. Exper., 32, 2002, 801–831. Torino, Italy, 2004. 60. J.A. Zachmann, A Framework for Information Sys- 75. M. Hori, T. Yoshida, Domain-oriented library of tems Architectures”, IBM Systems Journal, 26, 3, scheduling methods: design principles and real-life 1987, 276-292. applications, International Journal of Human-Com- 61. J.F. Sowa, J.A. Zachman, Extending and formalizing puter Studies, 49, 4, 1998, 601-626. the framework for information systems architecture”, 76. R. Mizoguchi, J. Vanwelkenhuysen, M. Ikeda, Task IBM Systems Journal, 31, 3, 1992, 590-616. ontology for reuse of problem solving knowledge - 62. R. Hilliard, IEEE Std 1471 and Beyond, SEI’s First towards very large knowledge bases, IOS Press, 1995, Architecture Representation Workshop, 16-17 46-57. January, 2001. 77. S.F. Smith, M.A. Becker, An ontology for constructing 63. M. Uschold, M. King, S. Moralee, Y. Zorgios, The scheduling systems, in Working Notes of 1997 Enterprise Ontology, M. Uschold and A. Tate (Eds.), AAAI Symposium on Ontological Engineering, The Knowledge Engineering Review, Vol. 13, 1998. AAAI Press, Palo Alto, CA, March 1997. 64. M. Grüninger, M.S. Fox, The Role of Competency 78. E. Motta, D. Rajpathak, Z. Zdrahal, R. Roy, The Question in Enterprise Engineering. in: Proc. of the Epistemology of Scheduling Problems, F. V. IFIP WG 5.7 Workshop on Benchmarking: Theory Harmelen, editor, Proceedings of the 15th Euro- and Practice, Trondheim, Norway, 1994. pean Conference on Artificial Intelligence, Lyon, 65. REA-ontology, http://www.msu.edu/user/mccarth4/ France, 2002, 215-219. rea-ontology/ 79. D. Rajpathak, E. Motta, Z. Zdrahal, R. Roy, A Ge- 66. M. Uschold, M. King, Towards a methodology for neric Library of Problem Solving Methods for building ontologies, in: Proc. of the workshop on Scheduling Applications, IEEE Transactions on Basic Ontological Issues in Knowledge Sharing, In- Knowledge and Data Engineering, Vol.18, 6, IEEE ternational Joint Conference on Artificial Intelli- Computer Society, 2006. gence (IJCAI’95), Montreal, 1995. 80. Simple Object Access Protocol (SOAP), http:// 67. A. Gómez-Pérez, A framework to verify knowledge www.w3.org/TR/SOAP sharing technology, expert systems with application, 81. Web Services Description Language (WSDL), http:/ Vol. 11, No. 4., 1996, 519-529. /www-106.ibm.com/ developerworks/library/w- 68. M. Grüninger, M.S. Fox, Methodology for the de- wsdl.html sign and evaluation of ontologies, in: Proc. of the 82. Universal Description Discovery and Integration Workshop on Basic Ontological Issues in Knowl- (UDDI), Technical White Paper, (September 6, edge Sharing, IJCAI-95, Montreal, Canada, 1995. 2000), http://www.uddi.org 69. IDEF, http://www.idef.com 83. P. Tolchinsky, J. Vazquez-Salceda, A. Staikopoulos, 70. PSL Ontology – Current Theories and Extensions, O. Cliffe, R. Riveret, W. Vasconcelos, H. Aldewereld, June 28, 1999, http://www.nist.gov/psl/pslontology/ State of the art, FP7-215890, ALIVE Project “Co- 71. Process Interchange Format (PIF), http://ccs.mit.edu/ ordination, Organisation and Model Driven Ap- pif proaches for Dynamic, Flexible, Robust Software 72. A. Gangemi, S. Borgo, C. Catenacci, J. Lehmann, and Services Engineering”, Deliverable D2.1, 2008,

28 I. Batchkova

http://www.ist-alive.eu numbered Report, LAP/GRAI, University of Bor- 84. U. Aßmann, S. Zschaler, G. Wagner, Ontologies, deaux I, France, 1992. Meta-models and the Model-Driven paradigm. in: 97. IDS, ARIS TOOLSET Documentation, IDS Prof. C. Calero, F. Ruiz, M. Piattini (Eds), Ontologies Scheer, Saarbrücken, Germany, 1995. for Software Engineering and Technologies, 98. AMICE, CIMOSA: Open System Architecture for Springer-Verlag, 2006, 49-102. CIM, 2nd revised and extended version. Springer- 85. G. Guizzardi, On ontology, ontologies, conceptuali- Verlag, Berlin, 1993. zations, modeling languages, and (meta) models, in: 99. G. Spur, K. Mertins, R. Jochem, Integrated Enter- Frontiers in Artifical Intelligence and Applications, prise Modelling, Beuth Verlag, Berlin, 1995. Databases and Information Systems IV, ISBN 978-1- 100. G. Booch, Object-Oriented Analysis and Design 58603-640-8, IOS Press, Amsterdam, 2007. with Applications, 2nd edition, The Benjamin/ 86. S. Brinkkemper, Formalisation of information sys- Cummings. Redwood City, CA, 1994. tems modelling, PhD Thesis, University of 101. J. Rumbaugh, M. Blaha, W. Premerlani, F. Eddy, Nijmegen, 1990. W. Lorensen, Object-Oriented Modeling and De- 87. E. Seidewitz, What Models Mean. IEEE Software, sign, Prentice Hall, Englewood Cliffs, NJ, 1991. 20, 5, 2003. 102. A. Kleppe, J. Warmer, W. Bast, MDA Explained: 88. F. Allilaire, J. Bézivin, F. Jouault, I. Kurtev, ATL: The Model Driven Architecture Practice and Eclipse Support for Model Transformation. In: Pro- Promise. Addison Wesley, 2003. OMG, MDA ceedings of the Eclipse Technology eXchange Work- Guide version 1.0.1. OMG document omg/2003- shop (eTX) at the ECOOP’2006 Conference, Nantes, 06-01, 2003. France, 2006. 103. OMG-SysML, The OMG Systems Modeling Lan- 89. E.D. Falkenberg, W. Hesse, P. Lindgreen, B.E. guage, http://omgsysml.org/index.htm, May, 2006. Nilsson, J.E. Han Oei, C. Rolland, R.K. Stamper, 104. T. Huynh, J. Osmundson, A system engineering F.J.M. van Assche, A.A. Verrijn-Stuart, K. Voss, A methodology for analysis systems of systems us- framework of information system concepts, The ing the System Modelling Language, Proceeding FRISCO report, 1998. of 1st Annual System of System Engineering Con- 90. P. Chen, The Entity-Relationship Model - toward a ference, Johnstown, PA, June 13-14, 2005. unified view of data, In: ACM Transactions on Da- 105. Rhapsody, http://modeling.telelogic.com/standards/ tabase Systems 1/1/1976 ACM-Press ISSN 0362- 106. K. Atanassov, Applications of Generalized Nets, 5915, 1976, 9-36. World Sciences, Singapore - New Jersey - Lon- 91. D. Ross, Applications and Extensions of SADT, in: don - Hong Kong, 1991. IEEE Computer, 18, 4, 1985, 25-34. 107. J.P. Bowen, M.G. Hinchey, Ten commandments 92. E. Yourdon, L.L. Constantine, Structured Design: revisited: a ten-year perspective on the industrial Fundamentals of a discipline of Computer Program application of formal methods, FMICS’05, Sep- and Systems Design, Prentice Hall, 1985. tember 5–6, Lisbon, Portugal, 2005. 93. T. De Marco, Structured Analysis and System Speci- 108. J.S. Ostroff, Formal Methods for the Specification fication, Prentice Hall, 1978. and Design of Real-Time Safety Critical Systems, 94. J. Martin, C. McClure, Structured Techniques for Journal of Systems and Software, Vol. 18, Num- Computing, Prentice Hall, Englewood Cliffs, 1986. ber 1, April, 1992, 33-60. 95. N.N., ICAM Integrated Computer Aided Manufac- 109. E.M. Clarke, J.M. Wing, Formal methods: sate of turing (ICAM) Architecture Part II, Volume IV- the art and future directions, ACM Computing Functional modeling Manual (IDEF0), Air Force Surveys, 28, 4, 1996, 626-643. Materials Laboratory, Wright-Patterson Air Force 110. J. Crow et al., Formal methods specification and Base, Ohio 45433, AFWAL-TR-81-4023, 1981. analysis guide book for the verification of software 96. G. Doumeingts, B. Vallespir, M. Zanettin, D. Chen, and computer systems, NASA guide book, 1997, GIM, GRAI integrated methodology - a method- http://eis.jpl.nasa.gov/quality/Formal_Methods ology for designing CIM systems, Version 1.0. Un- 111. F. Wang, Formal Verification of Timed Systems:

29 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

A Survey and Perspective, Proceedings of the IEEE, munication and software systems: real time (SFM- Vol. 92, No. 8, August, 2004, 1283-1305. RT), 2004, 1-24. 112. Ch. Attiogbé, Formal Methods Integration for Soft- 120. N. Lynch, R. Segala, F. Vaandrager, H. Weinberg, ware Development: Some Locks and Outlines, Re- Hybrid I/O automata, in: R. Alur, T. A. Henzinger, search Report, RR00.8, IRIN, Universite de E. D. Sontag (Eds), Hybrid Systems III, Verifica- Nantes, Juli, 2000, http://www.sciences.univ- tion and Control, Lecture Notes in Computer Sci- nantes.fr/irin/Vie/RR/indexGB.html ence, Springer Verlag, Vol. 1066, 1996, 496–510. 113. J.A. Bergstra, J.W. Klop, Algebra of Communicat- 121. R. Alur, C.N. Courcoubetis, T. Henzinger, P.-H. ing Processes with Abstraction, Journal of Theo- Ho, Hybrid automata: An algorithmic approach retical Computer Science, 37, 1985, 77-121. to the specification and verification of Hybrid sys- 114. C.A.R. Hoare, Communicating Sequential Pro- tems, Hybrid Systems, in: R.L. Grossman, A. cesses, Comm. ACM, 21, 1978, 666-677. Nerode, A. P. Ravn, H. Rischel (Eds.), Lecture 115. R. Milner, J. Parrow, J. Walker (1992). A calculus Notes in Computer Science 736, Springer Verlag, of mobile processes, I and II Information and Com- 1993, 209-229. putation, 100, 1, September 1992, 1-40, 41-77. 122. I.J. Hayes, C.B. Jones, J.E. Nicholls, Understand- 116. ISO 8807:1989, Information processing systems - ing the Differences between VDM and Z, Techni- Open System Interconnection - LOTOS - A for- cal Report UMCS-93-8-1, Department of Com- mal description technique based on the temporal puter Science: University of Manchester, http:// ordering of observational behaviour. Available www.cs.man.ac.uk/, 11 August, 1993. from ISO, http://www.iso.ch/. 123. T. Murata, Petri nets: Properties, analysis and appli- 117. P. Bremond-Gregoire, H. Ben-Abdallah, I. Lee, cations. Proceedings of IEEE, 77, 4, 1989, 541–580. Ordering processes in a real time process algebra, 124. F.D.J. Bowden, A brief survey and synthesis of the roles Proceeding of AMAST 3rd International Work- of time in Petri Nets, In: Mathematical and Computer shop on Real-Time Systems, March, 1996. Modelling, Vol. 31, 10-12, 2000, 55-68. 118. R. Alur, D.L. Dill, A theory of timed automata, 125. A. Pnueli, The temporal logic of programs, In Pro- Journal of theoretical computer science, 126, 2, ceedings of 18th IEEE Symposium on Founda- 1994, 183-235. tion of Computer Science, 1977, 46-57. 119. R. Alur, Decision problems for timed automata: A 126. E.M. Clarke, D. E. Long, K. L. McMillan, Compo- survey, in: Lecture Notes in Computer Science, sitional model checking, in: IEEE Press, Proceed- Vol.3185, Proc. of the 4th international school on ings of the Fourth Annual Symposium on Logic formal methods for the design of computer, com- in computer science, 1989, 353-362.

30 Journal of the University of ChemicalMohammad Technology Al-Shannag and Metallurgy, 47, 1, 2012, 31-36

ENERGY SAVINGS IN ALKYD-RESIN AND AQUEOUS-EMULSION PROCESSES USING A HEAT-INTEGRATED CYCLE

Mohammad Al-Shannag

Department of Chemical Engineering, Received 09 November 2011 Faculty of Engineering Technology, Accepted 12 January 2012 Al-Balqa Applied University, P.O. Box 15008, Marka 11134, Amman, Jordan.

ABSTRACT

In paint industries, synthesis of alkyd resins and aqueous emulsions consumes high energy levels required for occurrence of endothermic chemical reactions. In attempt to reduce energy consumption, the present work investigated energy integration of the heating cycles of aqueous-emulsion and alkyd-resin production lines that operate independently. The energy-integrated system between these two processes was built by removing the less efficient boiler of the aqueous- emulsion process and using the more efficient boiler of the alkyd resin for both processes. A shell-and-tube heat exchanger was applied to integrate both heating cycles. In order to maintain the production levels of the integrated processes, design calculations showed that the required surface area, based on the pipes outside diameter of the heat exchanger, must not be less than 22 m2. Furthermore, results indicated the positive role of the heat-integrated cycle in energy saving of around 12%. Keywords: energy integration, energy saving, alkyd resin, aqueous emulsion, heating cycles.

INTRODUCTION duction of alkyd resin and aqueous emulsion. As a case study, energy integration concept was applied on Nowadays where energy prices are increasing rap- the alkyd-resin and aqueous-emulsion production lines idly, due to population growth as well as improvement of of a Jordanian paints factory which is known as High living standards, energy conservation projects have be- Quality Company. In this company, the two processes come more important than ever [1-5]. In particular, paint operate independently without any heat recovery net- industries are energy intensive processes whereas the in- work. Each production line uses a boiler to generate cidence of energy cost is about 20% of the total produc- the utility streams required for accomplishing the en- tion cost [6, 7]. Therefore, the recovery of relatively small dothermic chemical reactions. Fig. 1 shows the alkyd- quantities of heat can accumulate to become sizeable resin process which uses one boiler to generate the util- energy savings [1]. The highly energy consuming nature ity stream of hot thermal oil which in turn enters the of such industrial processes is the key driving force for jackets of two reactors. This is to maintain the reactors improving its profitability and reducing the overall cost operating temperature in the range of 200-240oC. On of manufacturing [7, 8]. Hence, any attempt for energy the other hand, Fig. 2 illustrates the aqueous-emulsion conservation in the process goes a long way in many process in which a second boiler is used to generate a aspects. Chemical reactions involved in the synthesis of utility stream of hot water to feed the jacket of the emul- alkyd resin and emulsions are endothermic and thus sifier. This is to maintain its operating temperature in they need a continuous heat supply [8-12]. In this re- the range of 90-95oC. Metal coils are immersed inside gard, process system engineering plays an important role these reactors for cooling purposes of the products, in the energy-saving task [1, 4]. whereas liquid water comes from the cooling tower is The objective of the present study was to mini- used as a cooling medium. The processing time is about mize the energy consumption associated with the pro- 20 and 10 hours for completion of one batch of alkyd-

31 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

Mixer resin and aqueous-emulsion product, respectively. The batch is equivalent to 15 and 5 tons for alkyd resin and

10 Ton aqueous emulsion, respectively.

Pump 2 METHODOLOGY Oil boiler Resin reactor

Fig. 3 presents the heat-integrated cycles of the Mixer Pump 1 alkyd-resin and aqueous-emulsion processes. As shown 5 Ton in the figure the boiler of the aqueous-emulsion process Resin reactor was removed and a shell-and-tube heat exchanger was Pump 3 introduced. The energy integrated system was conducted according to the following stages. First, the operating Fig.1. Schematic diagram of the heating cycle for alkyd- conditions of the independent production lines shown resin process. in Figs. 1 and 2 were monitored. Second, the heat required for the aqueous-emulsion process was evaluated in order to know the heat load on the heat exchanger, thereby, the additional heat load on the boiler of alkyd- Mixer resin process. Third, efficiencies of boilers were assessed.

Water boiler Fourth, fluid flow rates and heat-transfer calculations 5 Ton Pump 4 were incorporated in the detailed design stage of the heat exchanger. Finally, energy consumption levels of Emulsifier Pump 5 both non-integrated and integrated processes were evaluated and compared to see the benefit of the adopted energy integration system. Fig. 2. Schematic diagram of the heating cycle for aqueous- emulsion process.

Mixer

10 Ton

Oil boiler Pump 2 Resin reactor

Mixer Pump 1 5 Ton

Pump 4 Resin reactor Pump 3

Mixer Heat exchanger 5 Ton

Emulsifier Pump 5 Water

Fig. 3. Schematic diagram of the heat-integrated cycle for alkyd-resin and aqueous-emulsion processes.

32 Mohammad Al-Shannag

77 m3 h-1 which is equivalent to mass flow rate of & = RESULTS AND DISCUSSION mo 16.47 kg s-1. The rate of heat transferred from alkyd- In the non-integrated process shown in Fig. 2, resin boiler to the thermal oil can be estimated using water which comes from the jacket of emulsifier will Eq. (1) to have a value of & = 912 kW. By using Eq. Qo enter the boiler at a temperature of around T = 90oC (2), the corresponding alkyd-resin boiler efficiency is 1 and it is heated to reach a temperature of around T = η= 0.78. Since the boiler efficiency of alkyd resin is 2 95oC. It was monitored that the heated water flows at a larger than the corresponding one of the aqueous-emul- volumetric rate of V& = 51.4 m3 h-1 which is equivalent sion process, the integrated-heat cycle was built by re- to a mass flow rate m& = 13.76 kg s-1 since water average moving the aqueous-emulsion boiler so that the heat density is about ρ = 963.7 kg m-3. The rate of heat duty of 289 kW will be added to the alkyd-resin boiler. w transferred from boiler to water can be calculated from This results in an overall heat rate of 1201 kW. While, the following sensible heat equation: the non-integrated processes consume 3937 litres of diesel per day, it can be demonstrated using Eq. (2) that & = & − (1) the amount of diesel consumed in the integrated pro- Q mC p (T2 T1 ) cess is 3446 litres per day. Therefore, a diesel saving of where C is the mean heat capacity of water which is around 12 % was achieved in the integrated energy sys- p estimated at the average temperature of around T = tem. Note that these processes operate 300 day per year. avg 92.5 oC to have a value of around C = 4203 J kg-1 oC-1). Hence, the amount of diesel consumed in the non-inte- p Substituting all known quantities into Eq. (1) gives a grated processes is 1181 m3 per year and it is 1034 m3 heat rate of around & = 289 kW. Furthermore, the per year for the integrated one. This leads to an annual Qw efficiency of the aqueous-emulsion boiler can be calcu- diesel saving of around 147 m3 (Fig. 4). lated using: Design computations of the heat exchanger In the integrated process, the heat required for Q& η = (2) emulsifier reactions is supplied via a stream of water & × V GHV which leaves the shell side of the heat exchanger and where Q& is the rate of heat transferred from boiler to the operating fluid (water) passes through the heating cycle, & is the volumetric rate of diesel consumed, V I ntegrated processes and GHV is the gross heating value of the fuel intro- Non-integrated processes duced to the boiler. The two boilers considered in this 1200 work use a diesel as fuel. The gross heating value of Diesel saving: 147 m3 per year -1 diesel is GHV= 38.6 MJ L . It was recorded that the 1000 aqueous-emulsion boiler consumes around 1320 litres 3 of diesel per day, i.e., & = 1.53x10-2 L s-1. Now, apply- V 800 ing Eq. (2) gives a boiler efficiency of about η= 0.49. On the other hand, the boiler of alkyd-resin pro- 600 Total Integrated cess, shown in Fig. 1, consumes around 2617 litres of processes diesel per day, i.e, & = 3.03×10-2 L s-1, to heat thermal V 400 Emulsion oil from T = 220oC to T = 240oC. At the average tem- process 1 2 perature of T = 230oC, the corresponding physical prop-

Annual ofm consumption diesel; Resin avg 200 erties of thermal oil are approximately as follows: the oil Process heat capacity is C = 2769 J kg-1 oC-1 and the oil density po 0 is ρ = 770 kg m-3.The thermal oil enters the jacket of the o 5 tons and 10 tons alkyd-resin reactors with a volumetric Fig. 4. Comparison between non-integrated and integrated 3 -1 flow rate of 32 and 45 m h , respectively. Hence, the oil processes of the Jordanian paint company (High Quality enters the alkyd-resin boiler with an overall flow rate of Company; Amman). Note that 1year = 300 operating days.

33 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

Water at 90 oC with flow rate of

Oil at 225 oC Oil at 230 oC with flow rate with flow rate of

Water at 95 oC with flow rate of

Fig. 5. Schematic diagram of the heat exchanger with its inlet and outlet streams. enters the jacket of the emulsifier at temperature of 1 1 1 1 1 ln(do / di ) T =95oC (Figs. 3 and 5). This stream leaves the jacket = + + + + + (5) s1 U A h A h A h A h A 2πk L at temperature T =90oC. Then, it is preheated in the o o p i s o do o di i steel s2 heat exchanger via a stream of thermal oil, coming from where h is the forced-convection heat-transfer p the heating cycle at temperature of T =230oC, to feed coefficient in the pipes-side, h is the forced-convection p1 s the pipes of the heat exchanger. This oil stream leaves heat-transfer coefficient in the shell side, A is the total i the heat exchanger at temperature of about T =225oC heat-transfer area based on the total inside diameter of p2 and it is re-circulated to the oil boiler. In order to en- the pipe, d , L is the pipe length, h is the fouling(dirty) i do hance the performance of heat exchanger, the shell-side coefficient for water, h is the fouling(dirty) coefficient di stream is chosen to be in a counter current flow with for thermal oil, and k is the thermal conductivity of steel the pipes-side stream [13]. the carbon steel selected to be the material of construction The basic design equation used to estimate the of both shell and pipes (k =43 W m-1 oC-1). steel heat-transfer surface area, A , based on the outside The thermal oil enters the pipes of the exchanger at o diameter of the tubes, d , is [14]: temperature T = 230oC and leaves at temperature, o p1 T =225oC. Physical properties of the thermal oil at the p2 & = ∆ (3) average temperature, T = 227.5oC, are : oil density is Q U o A o T m mp ρ =768.60 kg m-3, oil heat capaty is C =2780 J kg-1 oC-1, o po where ∆T is the mean temperature difference defined oil viscosity is µ =6.94x10-4 kg m-1 s-1 , and oil thermal m o as [14]: conductivity is k =0.124 W m-1 oC-1. The mass flow rate o of the oil, calculated using the sensible heat equation, is ∆ = − (4) & = 20.78 kg s-1 and the corresponding volumetric flow Tm Tm, p Tm,s mo rate is V& = 2.70x10-2 m3 s-1. Table 1 presents the dimen- where T is the mean temperature in the pipe-side sions of steel pipes and shell used to design the heat ex- m,p which is equal to T = 227.5oC and T is the mean changer. For one-pass of tubes with a square pitch of 0.0254 m,p m,s temperature in the shell-side which is equal to T = m, the number of tubes in this exchanger is n = 137 [14]. m,s 92.5oC. This leads to a mean temperature difference of The average velocity of the oil in each pipe is: ∆T =135 oC. The overall heat-transfer coefficient, U , m o & is based also on outside pipe diameter. It can be ob- 4V 2 (6) u = =1.6083m s -1 tained from the following equation [9]: p π 2 n di Table 1. Dimensions of steel pipes and shell used to design the heat exchanger. Side Nominal size Inside diameter Outside diameter (in) (m) (m)

Schedule 40 3/8 di = 0.0125 do = 0.0171 steel pipe

Schedule 40 16 Di = 0.381 Do = 0.406 steel shell

34 Mohammad Al-Shannag

The pipe-side Reynolds number, Re is: p ρ u D = w s eq = ρ Res 20792 (14) ou p di µ Re = = 22265 w p µ (7) o and Prandtl Number of water is about = . Eq. Prw 1.92 and the Prandtl number of oil is: (9) is now used to give a shell-side heat-transfer C µ coefficient of about h =1750.76W m-2 oC-1 . Pr = po o =15.56 s o k (8) The total outside surface area of the pipes is Sinceo the flow in pipes is turbulent (Re >10000), = π = 2 and the inside area p Ao ndo L 7.36m the following relation can be used to determine the heat- is = π = 2 . Note that these areas are cal- Ai ndi L 5.38m transfer coefficient, h , [13]: culated by setting L = 1 m, which will be corrected to p hd the actual one in order to obtain the required heat load. Nu = =0.027Re0.8 Pr 0.4 k (9) The fouling coefficients for both water and oil where Nu is the average Nusselt number. Hence, the are of about 5000 W m-2 oC-1 [14]. Substitution of the forced-convection heat-transfer coefficient in the pipes above values into Eq. (5) gives an overall heat-transfer is = -2 o -1 . coefficient of about = -2 o -1 and the h p 2414.23 W m C U o 98.53 W m C On the other hand, water enters the shell of the corresponding heat rate is & = ∆ = . In Q U o Ao Tm 97.90kW exchanger at temperature T = 95oC and leaves at tem- order to achieve the required heat load, the one-meter s1 perature, T = 90oC. At this average temperature, the length of the pipes must adjusted to be s2 physical properties of liquid water were approximated at L = 289 97.90 = 2.95m ≅ 3m . Therefore, the actual the average temperature, T = 92.5oC [13]. Water den- total outside surface area of the pipes is ms sity is ρ = 963.7 kg m-3, water heat capacity is C = 4203 = × = m2 and the inside area w pw Ao 3 7.36 22.1 J kg-1 oC-1 , water viscosity µ =3.09x10-4 kg m-1 s-1, water is = × = m2. Table 2 summarizes the speci- w Ai 5.38 3 16.1 thermal conuctivity is k =0.677 W m-1 oC-1. The mass flow fications of the exchanger and its corresponding operat- w rate of water in the shell side is & = 13.76 kg s-1 and its ing conditions. mw volumetric flow rate is of V& =51.4 m3 h-1. The cross- Pressure-drop and energy losses due to friction sectional area available for this flow is: in the heat exchanger The energy loss due to friction, E , and pressure π f = ()2 − 2 = 2 As Di nd o 0.08254 m (10) drop, − ∆p , across the pipes or the shell of the heat exchanger 4 can be estimated using the Fanning equation [14]: and the average velocity in shell side is: − ∆ = p = L 2 & E f 2 f u (15) = V = -1 ρ D us 0.1728 m s (11) As where f is the Fanning friction factor. For flow with Since water flows in the region enclosed between Reynolds number Re > 4000, it can be estimated using outside-surface of the pipes and the inside surface of the Colebrook equation [14]: the shell, equivalent diameter must be calculated to be 1  ε 1.256  used in estimation the she-side heat-transfer coefficient, = −log +  (16) 3.7D  h as [13]: f  Re f  s, where ε is the roughness of the pipe, for carbon steel it 4A = s Deq (12) is around ε = 0.046 mm . The corresponding head loss W is = , where g is the gravitational accelera- hL E f / g where W is the wetted perimeter defined as: tion. The power done by pump on the operating fluid is given by Power = m&E .Under the operating conditions = π ( + )= f W Di nd o 8.5568 m (13) and the geometrical specifications of the heat exchanger This gives an equivalent diameter of given in the previous section, the numerical solution of = . Then, the shell-side Reynolds number, Eq. (16) and the above definitions will give the results Deq 0.0386m Re is found as: presented in Table 3. It must be taken into consider- s 35 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

Table 2. Summary of the operating conditions and the levels. From process design point of view, the use of one configurations of the designed heat exchanger. boiler instead of two boilers, to perform the same heat

Tube side duty, reduces both the operating and maintenance costs. Operating fluid Thermal oil Inlet temperature 230 oC REFERENCES Outlet Temperature 225 oC Volumetric flow rate 97 m3 hr-1 1. J. Xiao-Ping,W. Fang, X. Shu-Guang, T. Xin-Sun, Tubes layout Square pitch of 0.0254 m Number of Tubes; n 137 Han Fang-Yu, Minimum energy consumption pro- 2 Outside surface area; Ao 22.1 m cess synthesis for energy saving, Resources, Conser- 2 Inside surface area; Ai 16.1 m vation and Recycling, 52, 2008, 1000–1005. Length of tubes, L 3 m Shell side 2. R. Adonyi, J. Romero, L. Puigjaner , F. Friedler, Incorpo- Operating fluid Water rating heat integration in batch process scheduling, Ap- Inlet temperature 90 oC plied Thermal Engineering, 23, 2003, 1743–62. Outlet temperature 95 oC 3. L.T. Biegler, I.E. Grossmann, A.W. Westerberg, Sys- 3 -1 Volumetric flow rate 45 m hr tematic methods of chemical process Design, 1st edn., Shell inside diameter 0.381 m Prentice Hall, New Jersey, 1997. Material of construction 4. R. M. Smith, Chemical Process Design and Integra- Carbon steel tion, 1st edn., John Wiley & Sons, New York, 2005. 5. M. Bagajewicz, H. Rodera, Energy savings in the total Table 3. Energy losses due to friction in the pipe–side site Heat integration across many plants, Computers and shell-side of the heat exchanger. and Chemical Engineering, 24, 2000, 1237-1242. 6. G.P.A. Turner, Introduction to Paint Chemistry and Parameter Pipe-side Shell-side Fanning friction factor, f 0.008 0.007 Principles of Paint Technology, 3rd edn., Chapman -1 Specific energy loss, Ef in J kg 1360.77 0.0325 and Hall, London, 1988. Pressure drop, - P in atm 13 0.0003 7. P.A. Lovell, M.S. El-Aasser, Emulsion polymeriza- Head loss, h , in m 138.71 0.0033 L st Power done by pump on fluid in hp 37.9 0.001 tion and emulsion polymers, 1 ed., Wiley New York, 1997. ation that the diameter used in the shell-side calcula- 8. D.S. Ogunniyi, T.E. Odetoye, Preparation and evalua- tions is the equivalent one and the heat exchanger con- tion of tobacco seed oil-modified alkyd resins, tains 137 pipes. It is found that pumps 4 and 5 of the Bioresource Technology, 99, 2008, 1300–1304. non-integrated aqueous-emulsion process shown in Fig. 9. J.P. Claverie, R. Soula, Catalytic polymerizations in aque- 2 are able to overcome these additional energy losses. ous medium, Prog. Polym. Sci., 28, 2003, 619–662. 10. S.C. Thickett, R.G. Gilbert, Emulsion polymeriza- CONCLUSIONS tion: State of the art in kinetics and mechanisms, Polymer 48, 2007, 6965-6991. To achieve a more efficient process with low 11. K.D. Weiss, Paint and coatings: A mature industry in energy consumption and operating cost, a systematic transition, Progress Polym. Sci., 22, 1997, 203–245. design tool is required in the process industries. The 12. J. Reimers, F.J. Schork, Robust nucleation in current work studied the synthesis strategy for polymerstabilized miniemulsion polymerization, J minimizing energy consumption through the integration Appl. Polym. Sci., 59, 833–41, 1996. of alkyd-resin and aqueous-emulsion processes. 13. J.P. Holman, Heat transfer, 9th edn., McGraw-Hill, Although the same heat load on the water boiler for New York, 2001. aqueous-emulsion process is applied on the oil boiler in 14. M.S. Peters, K.D. Timmerhaus, R. West, Plant De- the integration process, the total amount of diesel is re- sign and Economics for Chemical Engineers, 5th edn., duced by about 12 %; without reducing the production McGraw-Hill, New York, 2003.

36 Journal of the UniversityR. Boeva, of Chemical G. Radeva, Technology P. Hinkov, and E. Metallurgy, Hinkov 47, 1, 2012, 37-42

THERMAL AGEING OF DIFFERENT KINDS OF FIBER MATERIALS: A KINETIC STUDY

R. Boeva, G. Radeva, P. Hinkov, E. Hinkov

University of Chemical Technology and Metallurgy Received 14 September 2011 8 Kl. Ohridski, 1756 Sofia, Bulgaria Accepted 10 January 2012 E-mail: [email protected]

ABSTRACT

The aim of this work is to study the kinetics of the process of thermal ageing of fiber materials from the hard wood pulps with different degree of delignification. The investigation has been performed at the temperature of 105 oC, the brightness reversion is recorded from the 2nd hour till the 24st process hour. It is found that the ageing process is best described by the exponential kinetic equation which refers to processes taking place on uniformly inhomogeneous sur- faces. The coefficient of inhomogeneity and the values of the initial rate have been estimated. The current rate of the process decreases due to changes occurring in the chemical composition of cellulose and destruction of the structure of the fiber material. It was found that the bleached pulps aged more slowly. Keywords: thermal ageing, fiber materials, kinetics

IINTRODUCTION of the fiber materials starts with the oxidation of the phenol hydroxyl groups of the lignin. Oxidation leads The main purpose of bleaching is to increase the to formation of chromophors, which give the yellow brightness of the pulp. The reversion of brightness is the color of the pulp. process resulting in changes of the properties of fiber It is known that the bleached pulps age more materials under influence of light, air, humidity and heat slowly. The ageing reactions here are predominantly with the time [1]. The strongest effects of all types of related to the transformation of polysaccharides [7, 8]. influence have heat and light. The thermal ageing destructs Accelerated thermal ageing regenerates the chromo- the fiber materials due to the degradation reactions of phores destructed in the bleaching process [12]. cellulose chains. It causes decreasing of whiteness and The lignin and hemicelluloses reducing enzymes strength of the samples, and changes in their chemical – cellulase, laccases, xylanases and mannanases composition and proprieties [2, 3]. Lignin content is also significantly reduce the thermal ageing of the pulps [13]. one of the major factors ageing of pulp depends on. The Ageing is a complicated process which does not lignin is suspected of being a precursor in the formation obey formal kinetic principles. The rate of the changes of chromophores which lead to brightness reversion [4 - in fiber materials depends on the temperature and on 7]. That is why the different fiber materials have different the process duration. Some authors describe the behavior during ageing. The lignin content in high yield degradation reactions of cellulose chains due to ageing fiber materials, thermomechanical pulp and processes as reactions of the first or second order [13 - chemimechanical pulp is high and they are ageing faster 16]. The equations were derived for processes taking than bleached cellulose [8-11]. please on a homogeneous surface. But in a real From a chemical point of view, the heat-induced heterogeneous cellulose system, the influence of the ageing is an extraordinarily complex process, influenced physical structure of the surface has to be taken into by a large number of interacting factors. The yellowing account [17]. Investigations regarding the correlation

37 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012 between the kinetics of cellulose degradation and the the SemiCell during the first stage of bleaching. All changes in a wide range of macroscopic properties such concentrations are expressed as per cents (%) of the as brightness, Pc number or tensile strength have been absolute dry fiber material published [12, 18, 19]. Pulp 3 has not undergone any additional The aim of this work is to study the kinetics of bleaching. the process of thermal ageing of fiber materials from As a result, the brightness of the additionally hard wood pulps with different degree of delignification. bleached samples has increased as follows: CMP - W =61.1 %; SemiCell - W =56.1 %; Pulp 1 - W =65.9 0 0 0 % and Pulp 2 W =70.9 %. EXPERIMENTAL 0 Samples from all types of fiber materials (before This research studies fiber materials from and after additional bleaching) have undergone thermal different types of hard pulps, as follows: Chemical ageing into a thermal box at a temperature of 105oC. Mechanical Pulp (CMP), Semicellulose (SemiCell), The degree of brightness has been determined at time hardwood pulp obtained after enzyme hydrolysis (Pulp intervals ranging from 2 to 24 hour after the beginning 1), hardwood pulp obtained after alkali extraction and of the ageing process. O pretreatment (Pulp 2) and hardwood pulp taken after 2 O , ClO , alkali and ClO pretreatment - O-D -EO-D 2 2 2 o 1 RESULTS AND DISSCUSION (Pulp 3). All above-mentioned pulps samples were delivered by „Svilotcell” AD, Bulgaria. In order to investigate the influence of bleaching The amount of the lignin in the pulp stamps is on the ageing process, all fiber materials (untreated and minimal and in the process of bleaching all of it passes additionally treated) have been put to thermal ageing. into the bleaching solution. In the Chemical Mechanical The research tests were performed at a temperature of Pulp and in the Semicellulose the amount of the lignin 105 îÑ. The records of the brightness reduction were is much higher - as in the initial wood. In the process of tracked on the 2, 4, 6, 12 and 24th hour. bleaching the lignin can be only modified and the whole The results obtained for all fiber materials are amount remains in the initial material. Through the presented in Figs. 1 and 2. After 24 hours thermal age- existing quantitative methods for lignin determination, ing, the brightness of the samples has been reduced in the minimum reduction of lignin amount in bleaching various degrees depending on time. CMP and SemiCell could not be found. As seen from Figs. 1 and 2, in spite of the vari- The initial brightness (W ) of the fiber materials ous types of initial fiber materials, the brightness de- 0 was measured by spectrophotometer Gretag Magbeth creases with the thermal treatment duration. Spectroeye (ISO Brightness 2470) and was as follow: CMP The decrease of brightness of the same samples - W =42.5 %; SemiCell - W =34.3 %; Pulp 1 - W =44.1 of fiber materials can be used as a criterion for the effect 0 0 0 %, Pulp 2 - W =55.1 % and Pulp 3 - W =89.3 %. of bleaching on the ageing processes. With the aim to 0 0 All different types of fiber materials were subse- compare the decrease of brightness data for the initial quently subjected to additional bleaching. The CMP and and the additionally bleached fiber materials as result SemiCell have undergone a combined two-staged bleach- of thermal ageing, the variable α has been introduced. ing: stage I – bleaching with H O , stage II – bleaching It takes into account the relative brightness decreasing 2 2 with Rongalit C. The first stage of the bleaching is char- caused by the treatment: acterized by the following conditions: T=80oC, process α = − (1) Wo W duration 90 min, H O – 2 %, mass concentration – 10 where W is the initial value of the brightness of samples 2 2 0 %, additives: NaOH – 2 %, Na SiO - 5 %, MgSO - (%) and W is the current value of brightness 2 3 4 0,5 %, ÅDTA - 0,5 %. The conditions of the second corresponding to a definite reaction time, (%). stage of bleaching were: T=80oC, process duration 90 The kinetic studies have been performed to assess min, mass concentration – 5 %, additive Rongalit C in details the specificity and the mechanism of the process (NaHSO .CH O.2H O) – 2 %. Pulp 1 and Pulp 2 were investigated. The quantity (Eq. 1) is used as a kinetic 2 2 2 α bleached under the same conditions as the CMP and variable. Its increase in time means a decrease of the

38 R. Boeva, G. Radeva, P. Hinkov, E. Hinkov

Fig. 1. Decrease in brightness W, (%) for the initial pulps Fig. 2. Decrease in brightness W, (%) for the additionally with time. bleached pulps with time. brightness or increase of the degree of process. The re- were analytically examined for kinetic description of the sultant kinetic curves are presented in Figs. 3 and 4. process. The results obtained are presented in Table 1. Figures 3 and 4 show that the degree of ageing, It was found that the ageing process was best which takes into account the relative reduction in described by the exponential kinetic equation. It is valid samples’ brightness in time, increases with the time. The for processes taking place at uniformly inhomogeneous highest degree of thermal ageing (both for the initial surfaces and successfully applied in studies of enzymatic and the additionally bleached materials) was recorded hydrolysis processes [18 - 22]. According to the model for the Chemical Mechanical Pulps – 3 %. The lowest of uniformly inhomogeneous surfaces the active centers effect is recorded for the Pulp 3 - preliminary factory on the surface are distributed linearly depending on their bleached sample. energy and entropy. For this type of surface the rate of The various kinetic equations concerning diffusion, heterogeneous reactions decreases exponentially with the topochemical and other types of heterogeneous processes increasing of the process.

Table 1. Correlation coefficient R for kinetic equations describing different heterogeneous processes.

R Kinetic Type of material equations SemiCell SemiCell CMP CMP Pulp 1 Pulp 1 Pulp 2 Pulp 2 Pulp 3 add. bl add. bl add. bl add. bl First order equation 0,71431 0,81675 0,81954 0,79682 0,80611 0,79038 0,94376 0,8324 0,8475

Exponential equation 0,99199 0,99050 0,99330 0,99254 0,98597 0,98808 0,99747 0,98609 0,98932

Power equation 0,91546 0,95738 0,95610 0,94893 0,93262 0,93135 0,97231 0,96310 0,96901

Topochemical equation 0,91691 0,95552 0,95689 0,9470 0,93856 0,93766 0,98334 0,96585 0,97333

39 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

Table 2. Kinetic characteristics of the process: kinetic coefficient of inhomogeneity a and logarithm of initial rate v of the ageing 0 Fiber material Initial Add. bleached

αa lnv0 α a lnv0 SemiCell 2.0 1.98 0.8 1.02

CMP 1.5 1.55 0.5 1.99

Pulp 1 2.3 1.47 1.2 0.91

Pulp 2 2.2 1.75 0.9 0.53

Taking into consideration the current rate The slope of the lines obtained allows to deter- mine the coefficient of inhomogeneity α. Using to Eq. dα v = (3), the values of the coefficient of inhomogeneity a and dt the values of the initial rate v of the ageing process are 0 the exponential kinetic equation becomes: calculated. The data obtained is represented in Table 2. It is seen from the Table 2 that the initial rate of = −aα (2) the thermal aging process is different for different fiber v v0e where ν is the initial rate of the process and α is a materials. This is understandable, since of the diversity 0 kinetic coefficient of inhomogeneity. The coefficient a of the materials, the different samples have different accounts for the effect of the energy and entropy factors contents of lignin. It can be concluded that the samples on the process rate. with greater lignin content have higher initial rate v 0 The approximate integral form of Eq. (2) is and kinetic coefficient of inhomogeneity a, in comparison applied to describe the kinetics of delignification: to those of the additionally bleached materials. The calculation of current rate v values allows α = 1 à + 1 ln(vo ) ln t (3) further application of the exponential equation (2). The a a rate at any time at α=const is found to follow the Eq. All kinetic curves are linearized (α vs ln t) in (4) for the all samples studied: correspondence with Eq. (3). The linear dependencies 1 obtained are presented in Figs. 5 and 6. v = (4) at

Fig. 3. Kinetic curves for initial pulps. Fig. 4. Kinetic curves for additionally bleached pulps.

40 R. Boeva, G. Radeva, P. Hinkov, E. Hinkov

Fig. 5. Linear dependencies α vs lnt for initial pulps. Fig. 7. Dependence of the current rate v, h-1 on time for the initial pulps.

Fig. 6. Linear dependencies α vs lnt for additionally bleached Fig. 8. Dependence of the current rate v, h-1 on time for the pulps. additionally bleached pulps.

Figs. 7 and 8 show the current rate decrease with For both samples (initial and additionally time t for both types of samples. It is clear that bleached) decrease in the current rate is more obvious irrespective of the high initial rates, the processes start for the fiber material with a high content of lignin. This to slow down. The decreasing of the current rate of the may be associated with a higher content of process with the increase of α is according to the expo- chromophores, respectively. The increase of the nential kinetic equation (2). chromophores, as a result of the thermal ageing influence, Typical of the processes described by the leads to the decreasing of the brightness of the samples. exponential kinetic equation (2) is that they are fast only in the initial stages. The aging process also proceeds CONCLUSIONS faster in the initial stage with a high initial rate, v . Over 0 time, the current rate v decreases significantly for all For the purposes of this study the kinetics of the fiber materials. In thermal aging processes the current process of thermal ageing of different type fiber materials rate decreases due to changes occurring in the chemical was investigated. The kinetics of the ageing is explained by composition of the cellulose and the destruction of the an exponential kinetic equation applicable to the processes structure of the fiber material. taking place on uniformly inhomogeneous surfaces. The

41 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012 coefficient of inhomogeneity and the values of the initial paper exposed to various methods of accelerated rate are estimated. The current rate of the process decreases ageing. Part I, Journal of Cultural Heritage, 10, for all fiber materials. This decreasing is higher for the 2009, 222-231. fiber material with a high content of lignin 12. M. Sakhawy, Effect of bleaching sequence on paper ageing, Polymer Degradation and stability, 87, 2005, Acknowledgements 419-423. The authors would like to thank the Scientific Re- 13. A. Tran, Thermal ageing of hardwood kraft pulp search Centre of the UCTM, Sofia for the financial sup- bleached with a chlorine dioxide based sequence, Jour- port of these investigations. nal of pulp and paper Science, 28, 4, 2002, 115-121. 14. Sm Saad, A. El-Kholy, Kinetic investigation of ther- REFERENCES mal aging of unbleached egyptian kraft bagasse pa- per, Holzforschung, 34, 1980, 281-287. 1. C. Chirat, V. Chapelle, Heat- and light-induced bright- 15. X. Zou, T. Uesaka And N. Gurnagul: Prediction of ness reversion of bleached chemical pulps, J. Pulp paper permanence by accelerated aging:I: Kinetic Pap. Sc, 25, 6, 1999, 201-219. analysis of the aging process, II: Comparison of the 2. I. Forsskåhl, Brightness reversion, in Forest Products prediction with natural aging results, Cellulose, 3, Chemistry, Papermaking Science and Technology 4, 1996, 243-267. Series, Vol. 3, P.Stenius, Ed., Fapet Oy, Helsinki, 16. A. Sharples, Degradation of cellulose and its de- 2000, pp. 279-323. rivatives. A. Acid hydrolysis and alcoholysis. In: 3. I. Croon, S. Dillén, J-E. Olsson, Brightness reversion Bikales NM, L. Segal, Ed. Cellulose and cellulose of birch sulphate pulp, Svensk Papperstidn., 69, derivatives, v. V, Pt V. Wiley-Interscience, New 1966,139-154. York, 1971, pp 991–1006. 4. K.M.M. Eiras, J.L. Colodette, Investigation of euca- 17. H. Z. Ding, Z. D. Wang, On the degradation evolu- lyptus kraft pulp brightness stability, J. Pulp Pap. tion equations of cellulose, Cellulose, 15, 2008, Sci., 31, 1, 2005, 1-17. 205–224. 5. N.A. Jappe, O.A. Kaustinen, The carbonyl group and 18. E. Hristova, V. Lasheva, Z. Litovski, E. Valcheva, A. brightness reversion, TAPPI, 42, 1959, 206-232. Kuzmin, Kinetics of Ageing of Fiber Semi-manu- 6. J.L. Colodette, K.M.M. Eiras, R. Oliveira, G. factured Materials, Cellulose and paper, 3, 1990, Ventorim, The effect of wood supply and bleaching 20-23, (in Bulgarian). process on pulp brightness stability, 28th EUCEPA 19. S. Veleva, E. Valcheva, I. Valchev, G. Radeva, 2001. Conference, Lisbon. Proceedings, 2003, pp 65-73. Application of an exponential kinetic equation to 7. A. Granström, T. Eriksson, G. Gellerstedt, C. Rööst, the interaction of optical brighteners with pulp, Re- P. Larsson, Variables affecting the thermal yellowing act. Kinet. Catal. Lett., 72, 355-364. of TCF-bleached birch kraft pulps, Nordic Pulp Pap. 20. E. Valcheva, S. Veleva, G. Radeva, I. Valchev, Res. J., 16, 2001, 18-28. 2003, Enzyme action of the laccase-mediator sys- 8. J. Buchert, E. Bergnor, G. Lindblad, L. Viikari, M. tem in the pulp delignification process, React. Kinet. Ek, Significance of xylan and glucomannan in the Catal. Lett., 78 ; 183-191. brightness reversion of kraft pulps, TAPPI Journal, 21. G. Radeva, S. Veleva, I. Valchev, Forns of expo- 80, 6, 1997, 257-270. nential kinetic equation for chemical and adsorp- 9. V. Lasheva, M. Karsheva, Investigation on the ther- tion processes on uniformly inhomogeneous sur- mal ageing of printing paper, J. Univ. Chem. Tehnol. faces, J. Univ. Chem. Tehnol. Met., (Sofia), 41, 3, Met., (Sofia), 43, 4, 2008, 394-398. 2006, 365-370. 10. D. M. Fljate, Svoistva bumagi; M. Lesnaia promishle- 22. E. Valcheva, R. Boeva, E. Petkova, Z. Litovski, In- nost, 1986, 567-585, (in Rusian). vestigation of the kinetic of ageing of chemi- 11. B. Havlinava, S. Katuscak, M. Petrocova, A. Makova, mechnical pulp from polar wood, J. Univ. Chem. V. Brezova, A study of mechanical properties of Tehnol. Met., (Sofia), 38, 4, 2002, 41-50.

42 Journal of the UniversityH. Nurmesniemi, of Chemical K. Manskinen, Technology R. and Pöykiö, Metallurgy, O. Dahl 47, 1, 2012, 43-52

FOREST FERTILIZER PROPERTIES OF THE BOTTOM ASH AND FLY ASH FROM A LARGE-SIZED (115 MW) INDUSTRIAL POWER PLANT INCINERATING WOOD-BASED BIOMASS RESIDUES

H. Nurmesniemi1, K. Manskinen2, R. Pöykiö3, O. Dahl4

1 Stora Enso Oyj, Veitsiluoto Mill, FI- 94800 Kemi, Received 18 October 2011 Finland Accepted 10 January 2012 2 Stora Enso Oyj, Heinola Fluting Mill, FI-18101 Heinola, Finland 3 City of Kemi, Valtakatu 26, FI- 94100 Kemi, Finland, E-mail: [email protected] 4 Aalto University, School of Chemical Technology, Department of Forest Products Technology, P.O. Box 16300, FI-00076 Aalto, Finland

ABSTRACT

The bottom ash and fly ash investigated in this study originated from the large-sized (115 MW) combustion plant of a pulp and board mill complex. This combustion plant uses a bubbling fluidized bed (BFB) boiler for energy production. During the sampling period, when the bottom ash was sampled from the outlet of the boiler and the fly ash from the boiler‘s electrostatic precipitator (ESP), approximately 97% or the energy produced by the BFB boiler originated from the incineration of clean forest residues (i.e. bark, woodchips and sawdust), and 3% from the incineration of sludge from the primary clarifier of a wastewater treatment plant. The current Finnish legislation, which regulates the utilization of ash in forestry, sets minimum or maximum concentration limit values for Ca, the sum of phosphorous and potassium, as well as for As, Cd, Cr, Cu, Hg, Pb, Ni and Zn. Although the total element concentrations in fly ash were between 1.5 (K) and 40 (Cd) times higher than those in bottom ash, all element concentrations, as well as the sum of the phosphorous and potassium concentration in the bottom ash and fly ash, fulfilled the requirements of the Finnish fertilizer act for forest fertilizer, which came into force on 13 September 2011. Keywords: Ash, fertilizer, fluidized bed boiler, forest industry, waste.

INTRODUCTION sawdust, or wastewater treatment plant residuals (mill sludge) mixed together [1-3]. Energy generation from In the forest industry, energy derived from biomass biomass is an environmentally sound alternative to fossil burned in boilers, recovery furnaces and other fuels, since a particular advantage of the use of wood combustion devices is an important component of the bioenergy is that it could significantly reduce the total energy needed to produce various products. formation of CO emissions. In addition, due to 2 Biomass combustion in boilers typically involves the enormous demands for the reduction of greenhouse gas burning of wood residues and residuals from emissions in the area of the European Union (EU), 20% manufacturing operations that include bark, wood chips, by 2020, and the consequently increasing use of

43 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012 renewable energy sources, the utilization of carbon study formed part of a major project focusing on reduc- neutral fuels such as wood-based biomass will ing, reusing and recycling various ash materials origi- dramatically increase in the near future. However, a nating from Finnish pulp and paper mills incinerating disadvantage of energy production from biomass is the different fuel mixtures. Our previous studies have fo- generation of large amounts of ashes of different kinds cused on the physical and properties of the bottom and (i.e. bottom ash and fly ash), which are traditionally fly ash originating from industrial or municipal power disposed of at landfill sites. plants incinerating a mixture of forest residues (i.e. bark, In Finland, the forest industry, together with ther- wood chip and sawdust), peat or/and coal [11-13]. How- mal and electric power plants, produce 200 000 to ever, in this study we investigated for the first time the 300 000 tonnes of ash annually. Half of this ash is from physical and chemical properties of the bottom and fly wood and bark and the other half from fuel peat and ash originating from a Finnish pulp and board mill in- coal. Nearly 70 % of the energy generated by the power cinerating a mixture of forest residues (i.e. bark, wood plants in the Finnish forest industry originates from the chip and sawdust) and the sludge from the primary clari- burning of wood material [4-5]. Although the majority fier of a wastewater treatment plant, which is mostly of the ash is reused, part of the ash is still disposed of in composed of the wood-based fibres. Thus, in practice, the industry‘s own or in municipal landfills [6-7]. the bottom ash and fly ash investigated in this study However, the increasing costs of landfill disposal, both originated from the incineration of wood-based biom- in the form of waste tax or the deposit fee, as well as ass residues. difficulties in acquiring new sites for disposal and the stricter EU landfill directives that have been issued during EXPERIMENTAL the recent years, have been the driving forces in Bottom and fly ash sampling minimizing the amount of solid wastes arising for The bottom ash and fly ash investigated in this disposal [8]. Due to the above-mentioned facts, the need study originated from the large-sized (115 MW) for recycling options for ash has become increasingly combustion plant of a pulp and board mill complex. urgent in Finland, and there is therefore a growing trend This combustion plant uses a bubbling fluidized bed towards the utilization of industrial residues in the (BFB) boiler for energy production. During the sampling Finnish forest industry [9]. In the context of increasing period, when the bottom ash was sampled from the outlet bioenergy production, the recycling of ashes in forestry of the boiler and the fly ash from the boiler‘s electrostatic may solve the problem of ash disposal and reduce the precipitator (ESP), approximately 97 % of the energy necessity for commercial fertilizer application [10]. produced by the BFB boiler originated from the In Finland, national legislation regulates the incineration of clean forest residues (i.e. bark, woodchips utilization of ash in forestry. According to Finnish and sawdust), and 3 % from the incineration of sludge fertilizer legislation, which came into force in July 2009, from the primary clarifier of a wastewater treatment minimum or maximum limit values are set only for the plant. These forest residues originated from the wood total concentration of chloride (Cl), the sum of the handling plant and sawmill of the pulp and board mill phosphorous (P) and potassium (K) concentration, the complex investigated in this study, and they were calcium (Ca) concentration, and for the concentrations therefore clean residues. Approximately 60 % of the of certain heavy metals (i.e. As, Cd, Cr, Cu, Ni, Pb, Zn forest residues comprised pine (Pinus sylvestris) trees, and Hg) [11]. In this study, we concentrated on the most 30 % were willow (Salix) and 10 % birch (Betula important physical and chemical properties in the verrucosa and B. pubescens). bottom ash and fly ash originating from the power plant Sampling of the ashes was carried out over a pe- (115 MW) of a pulp and board mill located in Finland. riod of three days, and the individual samples (1 kg per We only focused on those physical and chemical sampling day) were combined to give one composite properties, nutrient or heavy metal concentrations for sample with a weight of 3 kg for both the bottom ash which minimum or maximum limit values are currently and fly ash. The sampling period represented normal set in Finland for ashes used as a forest fertilizer. This process operating conditions for the combustion plant,

44 H. Nurmesniemi, K. Manskinen, R. Pöykiö, O. Dahl

e.g. in terms of O content and temperature. The incin- determination, samples were dried overnight to constant 2 eration temperature in a bubbling fluidized bed boiler mass at 105°C in a drying oven (Termaks) according to is ca. 800°C, while in the electrostatic precipitator it is European standard SFS-EN 12880. Determination of the ca. 176°C. After sampling, the samples were stored in water-soluble phosphorous concentration in the ashes was plastic bags in a refrigerator (+4°C). A coning and quar- carried out according to European standard CEN/TS tering method [14] was repeatedly applied to reduce 15105. In this procedure, the ash is heated with water in the ash sample to a size suitable for conducting a closed container at 120°C for 1 hour. The concentra- laboratory analyses. tion of phosphorous in the extract was determined with a Thermo Fisher Scientific iCAP6500 Duo (United Determination of the mineral composition, pH, dry Kingdom) inductively coupled plasma optical emission matter content and neutralizing value of the ashes spectrometer (ICP-OES). A more comprehensive review To determine the mineralogical composition of of the standards, analytical methods and instrumentation the bottom ash and fly ash, X-ray diffractograms of is provided in our previous paper [13]. powdered samples were obtained with a Siemens D 5000 diffractometer (Siemens AG, Karlsruhe, Germany) using Determination of the total nutrient and heavy metal CuKá radiation. The scan was run from 2 to 80° (2- concentrations in the ashes theta scale), with increments of 0.02° and a counting To determine the total nutrient (Ca, Mg, P and K) time of 1.0 second per step. The operating conditions and total heavy metal concentrations (B, As, Cd, Cr, Cu, were 40 kV and 40 mA. Peak identification was carried Hg, Pb, Ni and Zn) in the ashes, the dried samples were out with the DIFFRACplus BASIC Evaluation Package digested with a mixture of HCl (3 mL) and HNO (9 3 PDFMaint 12 (Bruker axs, Germany) and the software mL) in a CEM Mars 5 microprocessor-controlled package ICDD PDF-2 Release 2006 (Pennsylvania, microwave oven with a CEM HP 500 Teflon vessels (CEM USA). The pH of the ashes was determined using a pH/ Corp., Matthews, USA) using USEPA method 3051A EC analyser equipped with a Thermo Orion Sure Flow [15]. The cooled solutions were transferred to 100 mL pH electrode (Turnhout, Belgium). The determination volumetric flasks and the solutions were diluted to volume of pH was carried out according to European standard with ultrapure water. The ultrapure water was generated SFS-EN 12880 at a solid-to-liquid (i.e. ultrapure H O) by an Elgastat Prima reverse osmosis and Elgastat Maxima 2 ratio of 1:5. Determination of the dry matter content of ion exchange water purification system (Elga Ltd; Bucks, the ashes was carried out according to European standard England). All reagents and acids were suprapure or pro SFS-EN 12880, in which a sample is dried overnight to analysis quality. Apart from Hg, the total heavy metal a constant mass in an oven at 105°C. The neutralizing concentrations in ashes were determined with a Thermo (liming effect) value was determined according to Fisher Scientific iCAP6500 Duo ICP-OES (United European standard SFS-EN 12945. A comprehensive Kingdom). The concentration of Hg in the ash was review of the standards, analytical methods and determined with a Perkin Elmer Aanalyst 700 cold-vapour instrumentation is presented in our previous paper [13]. AAS equipped with a Perkin Elmer FIAS 400 and AS 90 plus auto-sampler. A more comprehensive review of Determination of the total chloride and water-soluble the standards, analytical method and instrumentation is phosphorous concentrations of the ashes provided in our previous paper [13]. Determination of the total chloride concentra- tion of the ashes was carried out according to European RESULTS AND DISCUSSION standard CEN/TS 15289. In this procedure, the ash is extracted with HNO in a CEM Mars 5 microprocessor- Mineral composition, nutrients and physical and 3 controlled microwave oven with CEM HP 500 Teflon chemical properties of the ashes vessels (CEM Corp., Matthews, USA) at 120°C for 1 hour. The concentration of chloride in the extract was The XRD spectra in Figure 1(A,B) shows the determined with a Dionex DX500 ion chromatography similarities and differences between the mineral system (Dionex Corp., U.S.A.). Before the nutrient composition of bottom ash and fly ash, whereas the Table

45 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

1 presents the their mineral formulas and abundances Table 1), partly support the findings reported by van (%) in the ashes. It can be seen that only silicate minerals Herck and Vandecasteele [21], although we did not ob- (albite, biotite, microcline and quartz) existed in the serve any hydroxides in the ashes. The dry matter con- bottom ash. Although silicate minerals such as albite tent of the bottom ash and fly ash was very high (99.5%). and quartz were also present in the fly ash, this ash This is a disadvantage, as it may increase the amount of additionally contained carbonate minerals (calcite and dust generated during the handling of these residues. calcium magnesium carbonate), oxides (hematite and The acid neutralizing value (NV) is one of the lime) and anhydrite, which is a sulphate mineral. The most important indices in evaluating the liming effect existence of silicate minerals in the bottom ash value of the ash in relation to its use in forestry [22]. originating from the bubbling fluidized bed boiler is The capacity of a liming agent to neutralize soil acidity reasonable when considering that the bed material of a depends on the levels of soluble and hydrolysable bases fluidized bed boiler usually consists of silica sand [16]. such as oxides, hydroxides, carbonates and silicates. The silicate minerals in these ash fractions may also be Cations such as calcium, magnesium, and potassium are partly due to sand and soil particle contamination of the interactive counter-ions [23]. According to the forest residues during harvesting, transportation and neutralizing values in Table 2, the fly ash has a ca. 3.0 handling [17]. In addition, it may partly derive from the times higher capacity as a liming agent to neutralize decomposition of plant tissue-derived Si-based minerals soil acidity and act as a soil amendment agent than the during incineration, such as phytolith (SiO × nH O), bottom ash. This is reasonable in the light of the fact 2 2 which is often a structural component of plant tissues, that the Ca concentration in the fly ash (21.3%; d.w.) deposited between and within plant cells [18]. The was ca. 3.2 times higher than that in the bottom ash minerals of our ashes are partly consistent with the (6.8%; d.w.). The NV of 26.9% (Ca equivalents; d.w.) findings of Steenari and Lindqvist [17], who also observed for the fly ash indicates that ca. 1.4 tonnes of this residue anhydrite (CaSO ), calcite (CaCO ), lime (CaO) and would be required to replace 1 tonne of a commercially 4 3 quartz (SiO ) in wood ash, as we did. Furthermore, the ground limestone product produced by SMA Mineral 2 minerals of our ashes correspond with the findings of Ltd., the neutralizing value of which is 38 % (Ca equiva- Maedwad et al. [19], who observed albite (NaAl O ), lents; d.w.). In the bottom ash, the respective NV was 3 8 hematite (Fe O ) and quartz (SiO ), but also with the 8.7% (Ca equivalents; d.w.), indicating that ca. 4.4 tonnes 2 3 2 findings of Holmberg and Claesson [20], who observed of these ashes would correspondingly be required to microcline (KAlSi O ) in wood ash. replace 1 tonne of the above-mentioned limestone prod- 3 8 According to van Herck and Vandecasteele [21], uct. The Mg concentration of the fly ash also indicates the alkaline pH of the ash indicates that part of the that it is a better forest fertilizer than the bottom ash. dissolved metals occur as basic metal salts, oxides, The Mg concentration in the fly ash was ca. 4.3 times hydroxides and/or carbonates. Thus, the strongly alkaline higher than that in the bottom ash. In addition, the P pH value of 11.9 (bottom ash) and 12.8 (fly ash) in our and K concentrations were clearly higher in the fly ash ashes (see Table 2), as well as the minerals in them (see than those in the bottom ash, which also indicate that it

Table 1. The minerals, their formulas and abundances (%) in the bottom ash and fly ash.

Mineral Formula Bottom ash Fly ash (%) (%) Albite NaAl3O8 12.8 28.7 Anhydrite CaSO4 9.8 Biotite K(Fe,Mg)3AlSiO3O10(F,OH)2 8.5 Calcite CaCO3 18.3 Calcium Magnesium Carbonate CaMg(CO3)2 9.6 Hematite Fe2O3 3.6 Lime CaO 17.1 Microcline KAlSi3O8 44.8 Quartz SiO2 33.9 12.8

46 H. Nurmesniemi, K. Manskinen, R. Pöykiö, O. Dahl

is a better plant nutrient agent than the bottom ash. Compared to the total phosphorous content, i.e. Although current Finnish legislation on the use of ash 0.3% (d.w.) in the bottom ash and 1.5 % (d.w.) in the fly as a fertilizer in forestry does not set any limit value for ash, the water soluble phosphorous content in both ashes the chloride (Cl) concentration in ash [24], we have was negligible (< 0.01-%; d.w.). Water-soluble P rep- determined its concentration in this study because it resents the amount of P readily available to plants [27]. was also analysed in our previous studies [11, 13, 25]. According to Schiemenz and Eichler-Löbermann [28], The chloride content in both ash fractions was clearly the poor water solubility of phosphorous in wood ash lower than the maximum limit value (2.0%; d.w.) of the means that only a small part of P in ash is extractable Decree of 12/07 [26], which were in force until 13 and available for plants when ash is used as a forest September 2011. As a consequence of the Ca and Mg fertilizer. However, according to Moilanen et al. [29], concentrations, which are of great interest concerning the water-insoluble forms of phosphorous in forest the utilization of biomass ashes in forests, the fly ash is fertilizers minimise the risk of phosphorous leaching. a better plant nutrient and soil improvement agent that Augusto et al. [30] claimed that the solubility of phos- the bottom ash.

4000 A Bo tt om ash 3000 Quartz 2000

1000 0000 Microcline 9000 8000

(Counts) Lin 7000 6000 5000 Biotite Albite 4000 3000 2000 1000 0 5 10 20 30 40 50 60 70 80 2-Theta - Scale

B Fly Ash Calcite

4000 Ca-Mg-Carbonate Quartz

3000 Lin (Counts) Albite Hematite

2000 Lime Anhydrite

1000

0 5 10 20 30 40 50 60 70 80 2-Theta - Sca le

Fig. 1. XRD pattern for the bottom ash and fly ash.

47 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

Table 2. Physical and chemical properties of the bottom ash and fly ash with respect to the current Finnish limit values for ash used as a forest fertilizer, as well as the previous limit values that were used before 13 September 2011 [24, 26], and the literature values for the total concentrations of nutrients and heavy metals in Finnish wood ashes [36-41].

Parameter Unit Bottom Fly Limit value Limit value Literature value / metal ash ash after before for the total 13.09.2011 13.09.2011 concentration in Finnish wood ashes pH (1:5) 11.9 12.8

DMC (105 °C ) % 99.5 99.5 NV % (Ca; d.w.) 8.7 26.1 Cl % (d.w.) < 0.1 0.5 2.0 (max) Ca % (d.w.) 6.0 20.5 6.0 (min) 6.0 (min) 10 - 30 Mg % (d.w.) 0.6 2.6 1.0 – 3.0 P % (d.w.) 0.3 1.5 0.1 - 1.6 K % (d.w.) 2.6 3.9 3.0 - 6.0 P+K % (d.w.) 2.9 5.4 2.0 (min) 1.0 (min)

P(H2O) % (d.w.) < 0.01 < 0.01 B mg/kg (d.w.) 150 340 72 - 555 As mg/kg (d.w.) < 3.0 < 3.0 40 (max) 30 (max) 0.2 - 60

Cd mg/kg (d.w.) 0.3 12 25 (max) 17.5 (max) 0.4 - 40 Cr mg/kg (d.w.) 39 69 300 (max) 300 (max) 40 - 250 Cu mg/kg (d.w.) 18 100 700 (max) 700 (max) 15 - 300 Hg mg/kg (d.w.) < 0.1 < 0.1 1.0 (max) 1.0 (max) 0.01 - 1.0 Pb mg/kg (d.w.) < 3.0 33 150 (max) 150 (max) 15 - 1000 Ni mg/kg (d.w.) 16 38 150 (max) 150 (max) 20 - 250 Zn mg/kg (d.w.) 720 3360 4500 (max) 4500 (max) 15 - 10 000

phorous is most likely low because in wood ash it is (B) in ashes used as a forest fertilizer. However, their bound in compounds with a low solubility, such as apa- content has to be reported to the environmental au- tite (Ca (PO ) (OH,F,Cl)), or Fe-Al oxides, although our thorities when ash is used for such a purpose, and there- 5 4 3 mineral data (see Fig. 1 and Table 1) do not supported fore their concentrations are reported in this study. the presence of these minerals in the ashes investigated If we compare the physical and chemical prop- in this study. In this context it is also worth to note, erties of the ashes investigated in this study to those that according to Table 2, there are no minimum or investigated in our previous ones [13, 25], the main maximum limit values for the pH, dry matter content important differences are the total concentrations of (DMC), neutralizing value (NV), or the concentrations calcium (Ca) and magnesium (Mg) in the ashes. The of chloride (Cl), magnesium (Mg), phosphorous (P), total Ca and Mg concentrations in the bottom ash potassium (K), the sum of phosphorous and potassium investigated in this study were correspondingly ca. 3.2 (P+K), water soluble phosphorous (P(H 0)) or boron and 2.0 times higher than those in the ashes originated 2

48 H. Nurmesniemi, K. Manskinen, R. Pöykiö, O. Dahl

from a large-sized (32 MW) municipal district heating mines the final fate of volatile elements. Most of these plant incineration of peat fuel and clean forest residues species form compounds that condense on the surface such as stumps and sawdust [13]. However, the total Ca of particles in the flue gas, leading to the enrichment of concentration (20.5%; d.w.) in the fly ash investigated some elements in the fly ash fraction. This phenom- in this study was slightly lower than that (24.2%; d.w.) enon is well known and depends on many factors such in our previous one [13], although the total Mg concen- as the type of boiler, the fuel mix, tree species, soil type tration (2.6%; d.w.) in this study was slightly higher than and climate, and the efficiency of flue gas cleaning de- that (2.2%; d.w.) in the previous one. Furthermore, the vices, and has been reported elsewhere [8, 30, 33-35]. total Ca concentrations in the ashes investigated in this According to the results in Table 2, the concentrations study were ca. 1.8 times higher than in the ashes origi- of all elements in the bottom ash and fly ash were clearly nated from a 120 MW industrial power plant incinerat- lower that their current Finnish limit values and do not ing peat and wood residues [25]. restrict the utilization of these residues as a forest fer- tilizer. Total element concentrations in the ashes and their If we compare our total concentrations of potential as forest fertilizers nutrients and heavy metals in bottom ash and fly ash to Table 2 also presents the most relevant physi- those reported in literature (see Table 2), our cal and chemical properties of the bottom ash and fly observations for total concentrations are in relatively ash with respect to their current and previous Finnish good agreement with the concentrations of these forest limit values. In Finland, the new Decree on fer- compounds in wood ashes from Finnish power plants. tilizer products (24/11), which sets guidelines for ash The total concentrations of nutrients and heavy metals recycling and fertilizing in forestry, came into force in ashes depends, for instance, on the tree species, soil on 13 September 2011 [24], and substituted the De- type and climate where the trees are harvested, as well cree of 12/07 [26]. If we compare the current and pre- as the type of boiler, operating conditions, the fuel mix vious limit values for ash used as a forest fertilizer, and the efficiency of flue gas cleaning devices [33, 34]. according to the figures in Table 2, the new limit val- Therefore, the literature values [36-41] in Table 2 for ues ease the reuse of ash in forestry. Whereas the pre- the total concentrations of nutrients and heavy metals vious Decree of 12/07 [26] set a maximum limit value in wood ashes vary significantly. of 2.0% (d.w.) for Cl, the current Decree of 24/11 [24] no longer sets any maximum allowable concentration CONCLUSIONS for chloride. In addition, the maximum allowable As concentration of 40 mg/kg (d.w.) and cadmium con- Current Finnish legislation regulating the utili- centration of 25 mg/kg (d.w.) are higher in the current zation of ash in forestry only sets minimum or maxi- Decree than in the previous one. However, the maxi- mum concentration limit values for Ca, the sum of phos- mum sum of the phosphorous and potassium concen- phorous and potassium, as well as for As, Cd, Cr, Cu, tration (P+K) is stricter (min 2.0%; d.w.) in the cur- Hg, Pb, Ni and Zn. Although the total element concen- rent Decree than that (min 1.0%; d.w.) in the previous trations in fly ash were between 1.8 (Cr) and 110 (Pb) Decree. If we disregard elements whose concentra- times higher than those in bottom ash, all element con- tions were lower than the detection limits, the concen- centrations, as well as the sum of the phosphorous and trations of other elements in fly ash were between 1.5 potassium concentration in the bottom ash and fly ash (K) and 40 (Cd) times higher than those in bottom fulfilled the requirements of the Finnish fertilizer act ash. Compared to the bottom ash, the elevated ele- for forest fertilizer, which came into force on 13 Sep- ment concentrations in the fly ash indicate a high de- tember 2011. Therefore, current Finnish legislation con- gree of element volatilization [31, 32]. cerning fertilizer products does not restrict the utiliza- In addition to the element volatilization charac- tion of these residues as a forest fertilizer. As a conse- teristics, element retention in fly ash through other pro- quence of the Ca and Mg concentrations, which are of cesses (i.e. primarily the condensation process) deter- great interest for the utilization of biomass ashes in for-

49 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012 ests, the fly ash is a better plant nutrient and soil im- 11, 2007, 4118-4123. provement agent than the bottom ash. Compared to the 9. G. Watkins, M. Mäkelä, O. Dahl, Innovative use po- bottom ash, the fly ash investigated in this study had a tential of industrial residues from the steel, paper better capacity to act as a soil liming agent to neutralize and pulp industries – a preliminary study, Prig. Ind. soil acidity, because its neutralizing value (NV) was Ecol. Int. J., 7, 3, 2010, 185-204. 26.1% (Ca equivalents; d.w), whereas that of the bottom 10. K. Schiemenz, B. Eichler-Löbermann, Biomass ashes ash was 8.7% (Ca equivalents; d.w.). and their phosphorous fertilizing effect on different crops. Nutr. Cycl. Agroecosyst., 87, 3, 2010, 471- Acknowledgements 482. The authors wish to thank the technical staff of 11. R. Pöykiö, K. Manskinen, H. Nurmesniemi, O. Dahl, Suomen Ympäristöpalvelu Oy, who kindly conducted all Comparison of trace elements in bottom ash and fly the chemical analyses. Special thanks also to MSc Olli ash from a large-sized (77 MW) multi-fuel boiler at Taikina-Aho at the Institute of Electron Optics of the power plant of a fluting board mill, Finland, University of Oulu, for the XRD data, and Dr. Roy Siddall Energ. Explor. Exploit., 29, 3, 2011, 217-234. for the English language revision. 12. K. Manskinen, R. Pöykiö, H. Nurmesniemi, Com- parison of the total and fractionated heavy metal and REFERENCES sulphur concentrations in bottom ash and fly ash from a large-sized (120 MW) power plant of a flut- 1. A.V. Someshwar, J.P. Unvin, W.Thacker, L. Eppstein, ing board mill, Chemija, 22, 1, 2011, 46-56. B. Malmberg, Environmental aspects of wood resi- 13. H. Nurmesniemi, M. Mäkelä, R. Pöykiö, O. Dahl, due combustion in forest products industry boilers, Comparison of the fertilizer properties of ash frac- Tappi J., 10, 3, 2011, 27-34. tions from medium-sized (32 MW) and small-sized 2. C. Kunzler, Pulp and paper industry’s diverse organ- (6 MW) municipal district heating plant, Int. J. Eng. ics steam, Biocycle 42, 5, 2001, 30-33. Appl. Sci., 7, 2, 2011, 87-91. 3. D. Garvilesvu, Energy from biomass in pulp and pa- 14. R.W. Gerlach, D.E. Dobb, R.G. Raab, J.M. Nocerino, per mills. Environ. Eng. Manag., J, 7, 5, 2008, 537- Gy sampling theory in environmental studies, 1. As- 546. sessing soil splitting protocols, J. Chemometrics, 16, 4. M. Nieminen, M. Moilanen, S. Piirainen, Phospho- 7, 2002, 321-328. rus allocation in surface soil of two drained peatland 15. C. Yafa, J.G. Farmer, A comparative study of acid- forests following wood and peat ash application – extractable and total digestion method for the deter- Why effective adsorption on low sorptive soils? Silva mination of inorganic elements in peat material by Fenn., 41, 3, 2007, 395-407. inductively coupled-optical emission spectrometry, 5. J. Hytönen, Effects of wood, peat and coal ash fertili- Anal. Chim. Acta, 557, 1-2, 2006, 296-302. zation on Scots pine foliar nutrient concentrations 16. B.M. Steenari, S. Schelander, O. Lindqvist, Chemi- and growth on afforested former agricultural peat cal and leaching characteristics of ash from com- soils, Silva Fenn., 37, 2, 2003, 219-234. bustion of coal, peat and wood in a 12 MW CFB-a 6. A. Kepanen, M. Lodenius, E. Tulisalo, H. Hartikainen, comparative study, Fuel 78, 2, 1999, 249-258. Effects of different wood ashes on the solubility of 17. B.M. Steenari, O. Lindqvist, Fly ash characteristics cadmium in two boreal forest soils, Boreal Env. Res., in co-combustion of wood with coal, oil or peat, 10, 2, 2005, 135-143. Fuel, 78, 4, 1999, 479-488. 7. N. L. Meyers, M. J. Kopecky, Industrial wood ash as 18. G.S. Humphreys, D.M. Raven, R.J. Field, Wood- a soil amendment for crop production, Tappi J., 81, ash stone in Angophora costata (Gaertn.) J. Britt. 4, 1998, 123-130. following Sydney busfires, Aust. For., 67, 1, 2004, 8. M. Mahmoudkhani, T. Richards, H. Theliander, Sus- 39-43. tainable use of biofuel by recycling ash to forests: 19. A. Meawad, D. Bojinova, Y. Pelovski, Study on ele- Treatment of biofuel ash, Environ. Sci. Technol., 41, ments leaching from bottom ash of Enel Maritsa East

50 H. Nurmesniemi, K. Manskinen, R. Pöykiö, O. Dahl

3 thermal power plant in Bulgaria, J. Univ. Chem. 31. A. Demirbas, Heavy metal contents of fly ashes from Technol. Met. (Sofia), 45, 3, 2010, 275-282. selected biomass samples, Energy Source., 27, 13, 20. S.L. Holmberg, T. Claesson, Mineralogy of granu- 2005, 1269-1276. lated wood ash from a heating plant in Kalmar, Swe- 32. C-L. Lin, T.Y. Yeh, Heavy metals distribution char- den, Environ. Geol., 40, 7, 2001, 820-828. acteristics in different particle size of bottom ash 21. P. van Herck, C. Vandecasteele, Evaluation of the after agglomeration/defluidization at various fluidi- use of a sequential extraction procedure for the char- zation parameters, Biomass. Bioenerg., 34, 4, 2010, acterization and treatment of metal containing solid 428-437. waste, Waste Manage., 21, 8, 2001, 685-694. 33. B.M. Steenari, O. Lindqvist, Stabilisation of biofuel 22. F.S. Zhang, S. Yamasaki, M. Nanzyo, Waste ashes ashes for recycling to forest soil, Biomass. Bioenerg., for use in agricultural production: I. Liming effect 13, 1-2, 1997, 39-50. of plant nutrients and chemical characteristics of 34. M.W. Pasquini, M.J. Alezander, Chemical proper- some metals, Sci. Total Environ., 284, 1-3, 2002, ties of urban waste ash produced by open burning 215-225. on the Jose Plateau: implications for agriculture, Sci. 23. A. Saarsalmi, E. Mälkönen, S. Piirainen, Effects of Total Environ., 319, 1-3, 2004, 225-240. wood ash fertilization on forest soil chemical prop- 35. M. Moilanen, K. Silverberg, H. Hökkä, J. Issakainen, erties, Silva Fenn., 35, 3, 2001, 355-368. Wood ash as a fertilizer on drained mires-growth 24. Maa- ja metsätalousministeriön asetus lannoiteval- and foliar nutrients of Scots pine, Can. J. For. Res., misteista. Asetus nro 24/11 (Dnro 1784/14/2011; 35, 11, 2005, 2734-2742. 01.09.2011; in Finnish). Available from: http:// 36. H. Soininen, L. Mäkelä, A. Kyyhkynen, E. Muukkonen, www.finlex.fi/fi/laki/kokoelma/2011/20111015.pdf Biopolttoaineita käyttävien energiantuotantolaitosten (accessed on September 2011). tuhkien hyötykäyttö- ja logistiikkavirrat Itä-Suomessa, 25. K. Manskinen, H. Nurmesniemi, R. Pöykiö, O. Dahl, Tutkimuksia ja Raportteja Nro. 59, Mikkelin Comparison of the forest fertilizer properties of bot- Ammattikorkeakoulu, Mikkeli 2010, Kopijyvä Oy, tom ash and fly ash from the 120 MW power plant Mikkeli, Suomi (In Finnish; English Abstract: Utiliza- of a fluting board mill incinerating different fuel tion of biofuel consumption energy plants´ ash flows mixture, J. Int. Env. Appl. Sci., 5, 4, 2010, 526-537. and logistic flows in Eastern Finland]. 26. Maa- ja metsätalousministeriön asetus lannoiteval- 37. J. Jaakkola, Puu- ja turveperäisten tuhkien käyttö misteista. Asetus nro 12/07 (in Finnish). happamien vesien neutraloinnissa, Ympäristötekniikan 27. J.K. Fuhrman, H. Zhang, J.L. Schroder, R.L. Davis, kandidaattityö ja seminaari, Ympäristötekniikan M.E. Payton, Water-soluble phosphorus as affected koulutusohjelma, Teknillinen tiedekunta, by soil to extractant ratios, extraction times, and elec- Lappeenrannan Teknillinen Yliopisto, Lappeenranta trolyte, Comm. Soil Sci. Plan., 36, 7-8, 2005, 925- 2011, Suomi, Finland, (In Finnish). 935. 38. J. Perkiömäki, Wood ash use in coniferous forests – a 28. K. Schiemenz, B. Eichler-Löbermann, Biomass ashes soil microbiological study into the potential risk of cad- and their phosphorous fertilizing effect on different mium release. Ph.D.-Thesis, Finnish Forest Research crops, Nutr. Cycl. Agroecosyst., 87, 3, 2010, 471- Institute, Research Papers 917, 2004, Sävypaino, Espoo, 482. Finland. 29. M. Moilanen, P. Pietiläinen, J. Issakainen, Long- 39. P. Vesterinen, Wood as recycling – State of the Art in term effects of apatite and biotite on the nutrient Finland and Sweden, Technical Research Centre of status and stand growth of Scots pine (Pinus sylvestris Finland (VTT), Research Report PRO2/6107/03, 2003, L.) on drained peatlands, Suo, 56, 3, 2005, 115-128. VTT Processes, Finland. 30. L. Augusto, M.R. Bakker, C. Meredien, Wood ash 40. E. Ojala, Selvitys puu- ja turvetuhkan lannoite- sekä applications to temperature forest ecosystems – po- muusta hyötykäytöstä, Energiateollisuus ja Motiva, Fin- tential benefits and drawbacks, Plant Soil, 306, 1-2, land, (In Finnish: http://www.energia.fi; accessed on 2008, 181-198. December 2011).

51 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

41. M. Moilanen, J. Issakainen, Puu- ja turvetuhkien vaikutus ja puuston kasvuun, Metsätehon raportti Nro. 162, 2003, maaperään, metsäkasvillisuuden alkuainepitoisuuksiin Metsäteho Oy, Helsinki, Finland, (In Finnish).

52 Journal of the UniversityG. Yonkova, of Chemical A. Surleva, Technology T. Ginova-Stoyanova and Metallurgy, 47, 1, 2012, 53-58

TECHNOLOGY FOR PRODUCTION OF FLUORIDE ENRICHED BEER

G. Yonkova, A. Surleva, T. Ginova-Stoyanova

University of Chemical Technology and Metallurgy, Received 15 September 2011 8 “Kl. Ohridski” blvd, 1756 Sofia, Bulgaria Accepted 20 January 2012 E-mail: [email protected]

ABSTRACT

The production of protective foods has gained mush research attention recently due to their ability to reduce the level of radio contamination of human organism. Usually protective foods are widely used foods or drinks specifically designed to have decontaminative properties. The present work proposes a technology for production of fluoride enriched beer with radio protective properties. Naturally fluorinated mineral water was used as a source of fluoride. A mixture of fluorinated mineral and tap water in ratio 1:1 was used and pH adjustment with citric acid was applied during the mashing-in process. Fermentation of the fluoride enriched wort normally occurred. A 17 % decrease of fluoride content during the brewing process was observed. The fluoride enriched beer produced by the proposed technology met the physicochemical and sensory standards. Keywords: fluoride, wort, beer, decorporation, radio protection.

INTRODUCTION 3) altered products – obtained by replacing the exciting components with beneficial components; 4) Enhancing natural radio resistance of human or- enhanced commodities – changes in the raw commodi- ganism by food has gained much researcher‘s attention ties that altered the nutrient composition. in the functional food industry [1]. Functional food is a The term “protective food or drink” stands for natural or processed food or beverage that contains known a functional dietary product reducing the level of ra- biologically active compounds which when in defined dio contamination in two ways [4,5]: quantitative and qualitative amounts provides a clinically a) decorporation – cleaning of the body of ra- proven and documented health benefit [2]. Different types dionuclide contaminants; of functional food have been defined [3]: b) hindering the radionuclide contaminant ad- 1) fortified products – obtained by increasing sorption. Usually it is produced by fortifying a widely the content of exciting nutrients; used food by an appropriate decorporative agent. The 2) enriched products – obtained by adding new radioprotective agents minimise or prevent the dam- nutrients or components not normally found in a par- age from radiation exposure caused by nuclear power ticular food; facility, food radiation or other devices that releases

53 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012 radiation [6,7]. “Protective food and drinks” are effec- tive radio prophylaxis means due to the possibility for application to large groups of people for extended pe- riods of time. The annual consummation of beer in Bulgaria is around 70 litres per person and the drink is considered as the most consumed [8]. Moreover, it is known that the beer enhances the excretion function of kidneys and thus intensifies the elimination of toxins from the body. Hence, beer could be used in the radionuclide prophy- laxis if an appropriate agent could be found. Fluoride is well recognised as an essential micro- nutrient which participates in tooth and bone building. Fig. 2. Number of suspended cells in the control (n) and the Humans obtain fluoride mainly from their diet. All of the experimental (l) worts. fluoride available in water reaches the blood cycle. In con- trast, between 50 - 80 % of food fluoride is absorbed by human organism [9]. Furthermore, it was proved that fluoride lowers the body’s radioactive contamination and thus acts as a radionuclide decorporation agent espe- cially for cerium and strontium [10-12]. It was estab- lished that the decorporation activity of fluoride is higher if naturally fluorinated waters are used. The content of fluoride in mineral waters is between 5 and 25 mg dm- 3. Usually water is considered as mineral when it con- tains various soluble substances at high concentrations and its temperature in the spring is above 20oC [13,14]. Due to the biological activity of the micronutri- ent and the drink, fluoride containing beer can be ex- Fig. 3. Apparent extract of the control (n ) and the pected to have a radio protective activity. As water is a experimental (l) worts. basic beer ingredient, it was supposed that a higher ef- fect can be achieved if the fluoride containing mineral water was used in brewing. As we reported previously, mineral water spoiled the quality of the obtained wort and the mashing technology had to be modified [15]. The present study is aimed at developing a tech- nology for production of fluoride enriched beer with potential radio protective properties. As water influences the quality of the produced beer, the characteristics of the products have been carefully examined in every stage of the brewing process.

EXPERIMENTAL

Fluoride containing mineral water from the Fig. 1. pH decrease during the fermentation of the control Blagoevgrad region was used. Composition of used n l wort ( ) and the experimental wort ( ) obtained using tap mineral and tap water is presented in Table 1. A mix- and mineral water in ratio 1:1

54 G. Yonkova, A. Surleva, T. Ginova-Stoyanova

Fig. 4. Fluoride concentration at different stages of the brewing process. ture of mineral and tap water in ratio 1:1 was used. Yeast strain Saccharomyces carlsbergensis was The mashing-in pH adjusted by citric acid was 5.4. used in the fermentation at temperature 9oC. Yeast propa- Wort was produced from malt with 5.3 % of gation was 15 - 20x106 cells cm-3. Beer maturation took moisture; 78.7 % of grist extract; 12.6 % proteins and place at 4oC. Apparent extract, temperature, concentra- 142 mg dm-3 α-amine nitrogen. The saccharizification tion of suspended yeast cells and active acidity of the time was 10 - 15 min, colour was 3.5 EBC and viscosity wort were controlled daily during the fermentation pro- – 1.64 mPa s. cess. A single-step mash decoction method was used. Standard analytical methods for beer quality as- Sparing water was at the same composition as in the sessment were used according to the European Brew- mashing. The wort was produced by boiling with á-bitter ing Convention [16]. The fluoride concentration was acids containing hop products (90 mg dm-3). 70 % of hops determined at every stage of the brewing. A protocol were added 15 minutes, and 30 % of hops were added 45 for ion-selective potentiometric determination of fluo- minutes, after the start of the boiling process. The over- ride in mineral water, wort and beer has been devel- all procedure continued 105 minutes. oped [17]. Sample was acidified with HCl before analy-

Table 1. Chemical composition of the used water.

parameter tap water mineral water pH 6.7 8.5 calcium hardness, oH 1.0 0.1 magnesium hardness, oH 1.4 0.2 carbonate hardness, oH 2.6 13.0 residual alkali 2.2 12.9 Ca2+, mg dm-3 14.4 2.0 Mg2+, mg dm-3 3.9 1.6 F-, mg dm-3 0.3 11.5

55 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

Table 2. Physicochemical characteristics of the control wort and the fluoride enriched wort. parameter control worth experimental worth pH 5.25 5.58 colour, EBC 8.0 8.5 viscosity, mPa s 1.72 1.71

α -amine nitrogen, mg dm-3 157 136 final degree of fermentation, % 78.0 79.1 bitterness, BU dm-3 24.6 25.8 fluoride concentration, mg dm-3 0.11 5.0

Table 3. Characteristics of the fluoride enriched beer. parameter control beer fluoride enriched beer alcohol, % 2.84 2.96 apparent degree of fermentation, % 76.7 79.1 final degree of fermentation, % 78.0 79.1 colour, EBC 5.5 5.5 pH 4.1 4.4 bitterness, BU dm-3 16.2 17.0 fluoride concentration, mg dm-3 0.11 4.75

sis to dissolve any formed CaF . 15 ml of filtered and the wort, influencing the pH of the mash and the wort. 2 acidified sample was thoroughly mixed with 15 ml of The higher alkalinity of the mash hindered the activity mixture of 0.25 M Na Cit and 1 M NaCl. Natrium cit- of the enzymes and affected negatively starch hydroly- 3 rate was used as a masking and buffering agent and NaCl sis. The incomplete starch degradation lowered the fi- as an ionic strength adjustor. Potential of the potentio- nal degree of fermentation. Moreover, higher pH of the metric cell (fluoride selective and silver/silver chloride mash, of the sweet wort and of the sparing waters en- reference electrodes) was measured before and after stan- hance the extraction of some undesirable substances such dard addition of NaF. Method accuracy and precision as polyphenols from malt flakes, thus spoiling the qual- were 4% and 1,7 %, respectively. ity of wort and beer. We tested some technological so- lutions: (1) a mixture of tap and mineral water in ratio RESULTS AND DISCUSSION 1:1 was used and (2) pH and hardness adjustment was applied adding citric acid during mashing-in. The fluo- Mashing and wort production ride content of the mineral water was lowered after di- Our previous study had shown that higher re- lution with tap water, thus achieving the health levels sidual alkali and carbonate hardness of the mineral water recommended by authorities [18]. Additionally, a de- from the region of Blagoevgrad had a negative effect on crease of the fluoride content during the mashing pro-

56 G. Yonkova, A. Surleva, T. Ginova-Stoyanova

cess was observed. The final concentration of fluoride was 17 %. The final fluoride concentration in beer was in the cold wort was 5 mg dm-3. Table 2 presents the 4.75 mg dm-3 and completely met the requirements of physicochemical characteristics of the control wort, and European and national regulations [18]. the experimental wort obtained using mixture of tap The characteristics of produced beer are pre- and naturally fluorinated mineral water in ratio 1:1 at sented in Table 3. As can be seen from the results there pH of mash 5.4, adjusted with citric acid. As the results was not a significant differences between the fluoride show the wort obtained using the proposed technology enriched and the control beer, except for the fluoride has the same characteristics as the control wort. concentrations. The fluoride enriched beer produced using mineral water had balanced aroma and specified Fermentation and beer production flavour, bright colour and pleasant hop bitterness. The active acidity, the propagation of yeast, the fermentation rate and the cells settlement of the control CONCLUSIONS and the experimental wort were studied to evaluate the effect of mineral water on the fermentation process. A new technology for production of fluoride en- The changes in pH during fermentation of the riched beer using mineral water is proposed. Tap and experimental wort are presented in Fig. 1. Despite the mineral water in ratio 1:1 was used in the brewing. A pH adjusting of the acidity of the mash by citric acid, a slight adjustment by citric acid was applied during the mashing- difference in the initial pH of the experimental and the in process. No discrepancies in the fermentation process control samples has been observed. During the first three were observed. The final content of fluoride in the beer days of fermentation pH was decreasing sharply as a result meets the requirements of the Bulgarian regulation. The of chemical changes in the buffering system [19]. After produced beer has good taste and aroma. the fourth day the acidity was almost unchanged. The final pH of the experimental and the control samples Acknowledgements were 4.68 and 4.45, respectively. The authors wish to gratefully acknowledge the The number of the suspended yeast cells and the University of Chemical Technology and Metallurgy, Sofia, apparent extract of the experimental fluoride enriched Bulgaria, for the financial support of this research by Grant wort, and the control wort are presented in Figs. 2 and 3, 10750/2010 from the Science and Research Program. respectively. The extract’s content in the control sample decreased to 2,7 % till the end of the sixth day, but in REFERENCES the experimental wort the same value was obtained af- ter 7 days of fermentation. During the wort fermenta- 1. P. Jones, S. Jew, Functional food development: con- tion maximum cell growth was reached during the sec- cept to reality, Trend Food Sci. Techn., 2007, 387- ond day. The final yeast cells concentrations in the ex- 390. perimental and the control young beers were equal. 2. D. Martirosyan, (ed.), Functional foods and Chronic The characteristics of the experimental and the diseases, Science and Practice, Food Science Publ., control samples coincide well showing that the 2001. fermentation of fluoride enriched wort had normally 3. J. Spence, Challenges related to the composition of occurred. functional food, J. Food Comp. Anal., 19, 2006, S4- S6. Characteristics of the fluoride enriched beer 4. D. Maurya, T. Devasagayem, G. Ivair, Some novel The fluoride content was studied during the over- approaches for radioprotection and beneficial effect all brewing process. The results are presented in Fig. 4. of natural products, Ind. J. Exp. Biology, 44, 2006, A slight decrease in the fluoride content during the 93-114. mashing process was observed (down to 5 mg dm-3). 5. D. Taylor, G. Stradling, M. Henge-Napli, The scien- The next brewing procedures did not affect consider- tific background to decorporation, Radiat. Prot. ably the fluoride concentration. The overall decrease Dosim., 87, 2000,11-17.

57 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

6. N. Prasad, M. Srinivasan, K. Pagalenti, P. Venugopal, 12. Z. Paskalev, Decorporation effect of natural fluo- V. Menon, Protective effect of feluric acid on gamma- rine mineral waters in Bulgaria on some osteotropic radiation-induced micronuclei, dicentic aberration radionuclides, Rentgenol. Radiol., 21, 1982, 22-26. and lipid peroxidation in human lymphocytes, Mu- 13. BG Regulation 6 of 3 August 2004 on the Require- tation research, 603, 2006, 129-134. ments to bottled, natural, mineral, spring and table 7. A. Dixit, D. Bhatnagar, V. Kumar, D. Chawla, K. waters intended for human consumption, Bulgarian Fuhruddin, D. Bhatnagar, Antioxidant potential and State Gazette Nr 68 of 3 august 2004 and Bulgarian radioprotective effect of soy isoflavone against gamma State Gazette Nr 66 of 25 July 2008. irradiation induced oxidative stress, J. Functional 14. BDS 14947-80, Bulgarian natural mineral waters Foods, 2011, doi: xd.doi.org/10.1016/j.jff.2011.10.005. 15. G. Yonkova, V. Zhivkova, A. Surleva, Use of fluo- 8. www.pivovari.com ride containing mineral water in wort production, 9. L. Ivanova, Biological impact, prophylactic effect and St. Cerc. St. CICBIA, 12, 2011, 373-380. toxicity of flour, Lechitel, 2008, (in Bulgarian). 16. European Brewery Convention, www.european 10. Z. Paskalev, G. Kalajdziev, G. Kabakchieva, Using breweryconventions.org. some foods and drinks for radioprotection purposes, 17. L. Ilcheva, G. Kabakchieva, P. Bozadjiev, Z. Denchev, IInd Scientific Conference “Medicine of disasters”, Conventional and flow techniques for fluoride deter- 20-22 October, 1993, Sofia, Bulgaria, (in Bulgarian). mination in bioproducts, III National Conference of 11. L. Dyankov, G. Kalajdzhiev, Z. Paskalev, Dynamic Chemistry, 1998, Plovdiv, Bulgaria, (in Bulgarian). examination of callus and of 85-strontium kinetics 18. BG Regulation 23 about physiological standards in in the lower leg bones of rats with traumatic injuries nutrition, 2005. treated with fluorine mineral waters, Rentgenol. 19. I. Kapzev, Technology of beer and alcohol free bev- Radiol, 22, 1983, 55-59. erages, Zemizdat, Sofia, 1993, p. 146 (in Bulgarian).

58 Journal G.of theIvanova, University M. Momchilova, of Chemical N. Technology Rumyan, A. and Atanasova, Metallurgy, N. Georgieva47, 1, 2012, 59-62

EFFECT OF SACCHAROMYCES BOULARDII YEASTS ADDITION ON THE TASTE AND AROMATIC PROPERTIES OF KEFIR

G. Ivanova1, M. Momchilova1, N. Rumyan2, A. Atanasova2, N. Georgieva1

1 University of Chemical Technology and Metallurgy Received 28 November 2011 8 Kl. Ohridski, 1756 Sofia, Bulgaria Accepted 20 January 2012 2 Lactina, 101 Sofia str., 1720 Bankya, Bulgaria E-mail: [email protected], [email protected]

ABSTRACT

Kefir is a natural probiotic and contains vitamins, minerals, essential amino acids and easily digestible complete proteins that beneficially affect the human body. The subject of the present study is the improvement of gustatory and aromatic properties of Kefir culture. The effect of addition of Saccharomyces boulardii to Kefir culture was explored. The characteristics of Kefir were analyzed by determination of: pH, acidity, viscosity, diacetyl test, cell number of LAB and yeasts. Keywords: Kefir, Lactic acid bacteria (LAB), Saccharomyces boulardii.

INTRODUCTION microorganisms that help to overtake pathogenic organisms, repopulate the digestive tract and help in Kefir is a traditional popular Middle Eastern digestion [1]. beverage due to overall sense of health and well being In present study the Saccharomyces species, and when consumed [1]. It is a product of fermentation of precisely Saccharomyces boulardii species, are milk with kefir grains. Kefir grains look like pieces of investigated. Saccharomyces boulardii is a tropical strain coral or small clumps of cauliflower, which contain a of yeast first isolated from lychee and mangosteen fruit complex mixture of both bacteria (including various species by Henri Boulard [2]. It is related to Saccharomyces of lactobacilli, lactococci, leuconostocs and acetobacteria) cerevisiae, but differing in several taxonomic, metabolic, and yeasts (both lactose-fermenting and non-lactose- and genetic properties [3]. It is shown that S. boulardii fermenting) that are similar to friendly probiotic bacteria maintains and restores the natural flora in the large and found in yogurt. Kefir grains or mother cultures from grains the small intestine; so it is classified as a probiotic. S. are added to different types of milk [1]. The grains cause boulardii has been shown to be non-pathogenic, non- its fermentation that results in numerous components systemic (it remains in the gastrointestinal tract rather in the Kefir including lactic acid, acetic acid, CO , than spreading elsewhere in the body), and grows at the 2 alcohol (ethyl, alcohol) and aromatic compounds. That unusual high temperature of 37°C [4]. Saccharomyces provides kefir’s unique organoleptic characteristics: fizzy, is not a part of the naturally occurring guts flora like acid taste, tart and refreshing flavor. Kefir is also a some bacterial probiotics. Saccharomyces is resistant natural probiotic. Probiotics are foods containing live to stomach acids, bile and pancreatic juices, as it is bacteria, beneficial to health. Kefir contains live active tolerates varying pH levels, so it survival through the cultures of normal flora consisting strains of gut is greater than that of bacteria probiotics [5-7].

59 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

EXPERIMENTAL Description of qualitative reaction for existence of diacetyl: In this study 7 fermentations with use of Kefir In a small tube was putted 2 ml of the sample, 2 culture are carried out. The sample involved 6g Kefir ml 0.1 N NaOH and 0.2 g creatine. The sample must be grains in 0.400 L sterilized skimmed milk and 0.08 mg mixture well. The acetoine was oxidated to diacetyl. It Saccharomyces boulardii yeasts. Pure Kefir culture was was obtained colored by positive reaction. used as a control – without yeasts addition. The For viscosity measurements the viscosimeter cultivation conditions for all fermentations were (Brookfield DV-I Prime) was used. temperature 30°C, duration 16 hours and cooling for 4 For microscopic observations prepared smears hours down to 4°C. were stained with 2% solution of methylene blue for 20 Number of LAB (lactic acid bacteria) was min at room temperature and then washed with distilled determined in MRS (Lactobacillus MRS Agar) agar, water and dried at 37°C. Observations were made with M17 (M 17 agar acc. to TERZAGHI for microbiology) a light microscope Olympus CX31 (Japan). agar and PCA (Plate Count Agar) agar according to IDF qualities and Bulgarian standards [8]. Number of RESULTS AND DISCUSSION yeasts was determined in YGC (Chloramphenicol Yeast Glucose Agar) agar according to IDF qualities and In Fig. 1 the pH changes during the 7 fermenta- Bulgarian standards [9-11]. tions are presented. The duration of each fermentations After the 7th fermentation the fermentation broths was 16 hours with 4 hours break for cooling down to obtained from control and sample were lyophilized. Each 4°C. The observed variation is due to addition of some of the lyophilized culture liquids was taken in 18 mg aromatic cultures – strains of Str. thermophilus. The and placed in 0.400 L skimmed milk. The lyophilization normal value of Kefir’s pH is about 4.40. At the beginning was carried out at -50°C under vacuum on laboratory pH started to decrease and that was the reason to add lyophilizator (Martin Criss). some cultures with aromatic properties after 3rd and 5th For pH measurement during the experiment pH fermentations to keep the pH in the optimal range. meter (Metler Toledo) was used. The results of measurement of active acidity are Measurement of active acidity: shown in Fig. 2. The acidity also varies after 3rd With pipette were measured 10 cm3 of the fermentation due to addition of cultures with aromatic sample, 20 cm3 distilled water, 3 drops of properties. Qualitative reaction for existence of diacetyl phenolphthalein and the mixture was shacked. The was done for all fermentations. The obtained results are sample was measured with titration with 0.1 N NaOH shown in Table 1. This qualitative reaction shows aro- until appearance of slight pink color. This color must matic properties (existence of diacetyl) of Contol and keep 1 min.

Fig. 1. Variation of pH during fermentation.

60 G. Ivanova, M. Momchilova, N. Rumyan, A. Atanasova, N. Georgieva

Fig. 2. Variation of titratable acidity during fermentation

Fig. 3. Microscopic picture of kefir fermentation product after 1st (a) and after fermentation 7th (b)

Table 1. Results of qualitative reaction for existence of diacetyl.

Culture Time of appearance for each of 7th fermentations [min] Average time 1 2 3 4 5 6 7 Control 3:30 3:00 2:00 2:30 4:00 3:00 4:00 3:08 Sample 3:30 4:00 2:00 3:00 4:30 3:30 4:30 3:48

Table 2. Cell number of LAB and yeasts. Culture Cell number [CFU/mg] Cell number of LAB Cell number of yeasts After 1st After 7th After 1st After 7th fermentation fermentation fermentation fermentation Control 2.5 x 108 2.5 x 109 5.5 x 105 1.1 x 105 Sample 2.5 x 108 2.0 xx 10108 9.90 5 14.8 x 106

Sample. Diacetyl was appeared in Sample and Control of LAB and yeasts count are presented in Table 2. Double at the same time. increase in the number of yeasts was detected by adding Cell number of LAB and yeasts was monitored at of Saccharomyces boulardii strain. During the experi- the beginning and in the end of the experiments. Results ment microscopic picture was traced. Microscopic im-

61 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

Fig. 4. Comparison between freeze-dried Control and freeze-dried Sample ages of kefir culture are shown in Fig. 3 - after the first REFERENCES fermentation (Fig. 3a) and after the seventh fermenta- 1. S. Otes, O. Cagindi. Kefir: A Probiotic Dairy-Composi- tion (Fig. 3b). After the last fermentation the number tion, Nutritional and Therapeutic Aspects. Pakistan of yeast has doubled while the number of lactic acid Journal of Nutrition, 2, 2, 2003, 54-59. bacteria remained the same. It was measured pH, ac- 2. J-P. Buts. Twenty-Five Years of Research on Saccha- tive acidity and viscosity of the Kefir, obtained from romyces boulardii Trophic Effects: Updates and Per- control and sample’s lyophilization. The results show spectives, Dig. Dis. Sci., 54, 2008, 15-18. that the values of each parameter are in the optimal 3. J.Y. Malgoire, S. Bertout, F. Renaud, J.M. Bastide, M. range for this kind of Kefir product (Fig. 4). This shows Mallié, Typing of Saccharomyces cerevisiae clinical that the industrial Kefir product sample doesn’t differ strains by using microsatellite sequence polymorphism, significantly from the control. The data obtained in J. Clin. Microbiol., 43, 3, 2005, 1133–7. this study can be useful in industrial practical to ad- 4. L. McFarland, P. Bernasconi Saccharomyces boulardii: just. a review of an innovative biotherapeutic agent, Microb. Ecol. Health. Dis. 6, 4, 1993, 157–71. CONCLUSIONS 5. D. Czerucka, T. Piche, P. Rampal. Review article: Yeast as probiotics - Saccharomyces boulardii. Ali- In conclusion some improvement of gustatory ment. Pharmacol. Ther., 26, 6, 2007, 767-778. properties of Kefir culture was established by addition 6. A. Sezer, U. Usta, I. Cicin. The effect of Saccharomyces of the Saccharomyces boulardii yeasts. Double increase boulardii on reducing irinotecan-induced intestinal mu- in the number of yeasts was detected by adding of cositis and diarrhea. Med. Oncol., 26, 2009, 350-357. Saccharomyces boulardii strain. This was established in 7. M. Guslandi, P. Giollo, P. A. Testoni. A pilot trial of Saccha- the microscopic picture. Positive influence was found romyces boulardii in ulcerative colitis. European Journal of by addition of aromatic cultures - strains of Str. Gastroenterology and Hepatology, 15, 2003, 697-698. thermophilus, affecting the sample for the presence of 8. Å. Simova, Z. Spasov, Handbook of microbiology of milk diacetyl in Kefir products. and milk products, Nova Print, Plovdiv, 2007, p. 192, (in Bulgarian). Acknowledgements 9. FIL-IDF 146:1991; 149A:1997; Bulgarian Standard No 10943-91. The present work is funded and supported by 10. FIL-IDF 94 B: 1990. Project No 10860 of University of Chemical Technology 11. H. Chomakov, N. Georgieva, Handbook of microbiol- and Metallurgy Science Fund “Research Investigations”. ogy, UCTM, Sofia, 2004, p. 86.

62 Rachid Souilah, Badreddine Belhadi, Nadia Boudries, Journal of the UniversityDjaafar of Chemical Djabali, Technology Boubekeur and Nadjemi Metallurgy, 47, 1, 2012, 63-68

PRETREATMENTS EFFECT OF SORGHUM (Bicolor (L.) Moench) AND MILLET (Pennisetum Glaucum) FLOURS ON THE IN VITRO STARCH DIGESTIBILITY

Rachid Souilah1, Badreddine Belhadi1, Nadia Boudries1,2, Djaafar Djabali1,2,

Boubekeur Nadjemi1

1Laboratoire de recherche sur les produits bioactifs Received 20 January 2009 et valorisation de la biomasse, Accepted 10 January 2012 Département de Chimie, Ecole Normale Supérieure, BP 92 Kouba, Alger (16050) Algérie E-mail: [email protected] 2Département de Génie de l’Environnement, Ecole Nationale Polytechnique, 10 Avenue Hassan Badi, El-Harrach, Alger Algérie

ABSTRACT

The starch digestibility of sorghum and millet flours was studied using 1 % suspensions of sorghum and millet flours. The two suspensions were exposed to an enzymatic hydrolysis by glucoamylase. After two hours the digested starch was measured. At these conditions the starch digestibility of both flours is approximately 36.4 % for sorghum and 36.8 % for millet. After that the two flours were pretreated with hexane to remove some lipids. The starch digestibility was estimated as 36.6% for sorghum and 36.2% for millet. Then some proteins were removed using a pretreated flours with the pepsin enzyme and the starch digestibility was estimated as 56.6 % for millet and 55.4 % for sorghum. After that the two suspensions were treated with hexane and pepsin. An increase of the starch digestibility was observed in both samples, 87.6 % for sorghum and 84.5 % for millet. Finally, after starch isolation, the starch digestibility increased to 89.5 % for sorghum and 89.7 % for millet. Keywords: glucoamylase, sorghum, millet, flours, digestibility, starch, pretreatments.

INTRODUCTION and substrats. The sorghum grain displays a low digestibility of starch due to the endosperm proteins [5] Sorghum and millet are an important source of and the presence of fibres, polyphenol compounds, dietary energy and the starch represents the major source resistant starch and Kafirins protein [6,7]. Zhang and in these cereals that are well adapted to African and Hamaker demonstrated the effect of protein of cooked Asian semi-arid and subtropical agronomic conditions sorghum flour on the alpha amylase digestibility [8]. [1]. Amylases are the enzymes that hydrolyse starch Elkhalifa at al. detected the increasing effect of contained in foods. The rate of starch digestibility can fermentation and the decreasing effect of resistant starch be estimated in vitro [2]. The starch digestibility of both on the sorghum starch digestibility [9]. The digestibility cereals depends on the presence of reducing agents such of foods (wheat, rice, etc.) was inversely related to the as proteins, lipids, fiber and the chemical and physical proteins, fats and the total dietary fiber contained in the characteristics of cereal starch [3]. The effect of amylase- foods [10]. lipid complex on the gelatinization of starch [4], as The main objective of this study is to evaluate well as the reducing impact of proteins on digesting the starch digestibility of sorghum and millet grain enzymes of starch, reduce the contact between enzyme cereals, grown in the south of Algeria, and to demonstrate

63 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012 the influence of some factors on the kinetics of the hy- was added to pH 2.5 for 2 h at 50°C. By the end of the drolysis. reaction the mixture was filtered and the solid residue Our research progressed as follows: first, several was dried in the drying oven at 40°C. The residue was samples of sorghum and millet flours were prepared. used for the starch digestibility determination. Then the samples were exposed to different pretreatments and finally the starch digestibility by Sample preparation of lipid isolated from sorghum glucoamylase enzyme was measured. and millet flours (sample 3) The method used for a total lipid extraction was EXPERIMENTAL described by [15]. Quantity of 2 to 5 g of flour was put in a sample tube used in a Soxhlet extractor of the fat, Substrates and enzyme were shaken for 20 min, each time 10 ml of hexane was Sorghum (Sorghum bicolor (L.) Moench) and added in 35°C. The solid material of the tube was millet (Pennisetum glaucun) grains cultivated in Tidikelt, extracted and dried at 40°C overnight to evaporate the a hyper arid region situated in the south of Algeria, solvent. were purchased from the 2005 harvest. All chemical products were Sigma Chemical Co. (St. Louis, MO) and Sample preparation of protein and lipid isolated from Merck certified grade. The enzyme amyloglucosidase sorghum and millet flours (sample 4) from Rhizopus Mold A7255, was Sigma, 23.000 U/g The protein and lipid were removed by the solid of enzyme. Enzyme pepsin was prolabo, 94120, methods mentioned previously. 100 U/mg. Sample preparation of starch isolated (sample 5) Chemical composition of sorghum and millet flours Starch was isolated from the sorghum and millet We relied on several methods for the estimation cultivars by alkali extraction as proposed in [16,17] with of the chemical composition. Ash was measured by NF few modifications. 10 g of grains were steeped in 20 ml 03-720 (1981) method described in [11]. Soluble sugars of 0.25 % (w/v) NaOH for 48 h. They were washed and were determined by extraction with alcohol and then crushed using a sample mill IKA labotechnik A10. measured with phenol-sulphuric acid method [12]. Pro- The suspensions were passed through a set of sieves (80, tein content estimation was based on the Kjeldhal method 100 ìm). The filtrates were centrifuged (30.000 tour/min ISI 24-1. The total lipid estimation was based on the during 10 min). The layer of the residual proteins was method described in [11], ISO 3947. Insoluble fibres es- scraped each time. The extract was then dried at 40°C timation was done by Van Soest method [12]. The total overnight. humidity was estimated according to ISI 01-1. The total starch content in the flour samples was measured by Digestibility test the enzymatic method [13]. The enzymatic digestibility of substrats was tested by the method described in [18]. A commercial Flour sample preparation (sample 1) crystalline glucoamylase enzyme of Rhizopus mold, Flour samples from sorghum and millet grains which has high affinity to raw starch, was utilized for were prepared by milling through a 500 ìm screen Retsch the digestion experiments. A reaction mixture consisting AS200 type using a sample mill IKA labotechnik A10. 0.13 g equivalent of starch and 6 ml of distilled water in the tubes was placed in a water bath at 100°C for 10 Sample preparation of protein isolated from sorghum min to obtain the starch suspension. After that the and millet flours (sample 2) samples were cooled to 40°C for 10 min, then 3.5 ml of The method used for the pepsin pretreatment of 100 mM acetate buffer (pH 5.0) and 3.5 ml of sorghum and millet is given in [14]. Flour samples glucoamylase solution (5 unit) were added. The samples (equivalent to 2.6 g of starch) were incubated in pepsin were incubated at 40°C for 2 h with stirring. After solution (0.4 g of enzyme in 100 ml of distilled water) digestion, surplus starch was removed by centrifuge in 250 ml batch reactor. Then 0.1 N hydrochloric acid (4350 tour/min during 10 min), and the contained

64 Rachid Souilah, Badreddine Belhadi, Nadia Boudries, Djaafar Djabali, Boubekeur Nadjemi

glucose of the filtrate was analyzed using the glucose oxydase-peroxidase method [19]. The percentage of starch is counted by multiplying the glucose percentage 80 with the factor 0.9. The reported values are the means of triplicate measurements. 60

Statistical Analyses 40 Each experiment for digestibility test included three replicates. The difference of means were index digestibility (%) 20 determined using Student’s t-test. Values of P < 0.05 were considered statistically significant. The results were 0 evaluated for significance by analysis with SPSS V.13. SR1 SR2 SR3 SR4 SR5 sorghum flours with different treatments RESULTS AND DISCUSSION Fig. 1. The index digestibility of sorghum flours at different Chemical Composition of sorghum and millet flours treatments. Table 1 demonstrates the chemical composition of sorghum and millet flours. The estimated values for the chemical composition of sorghum and millet flours shown in Table 1 are 65 % for total starch, 11-13 % for protein and 4-6 % for lipid. By comparing the results of 80 Barikomo et al. for sorghum and millet cultivated in Mali, the percentage of protein is 10.3 % for sorghum and 7.2 % for millet, the starch content is 65.6 % for 60 millet and 73.5% for sorghum [20]. These results indi- cate that the proteins and lipids have been affected the 40 starch digestibility of sorghum and millet by glucoamylase. index digestibility (%) 20

Digestibility test of samples 0 The ability of the amyloglucosidase to digest dif- ML1 ML 2 ML 3 ML 4 ML 5 ferent sorghum and millet flours was studied using millet flours with different treatments different treatments (flours, flours treated by pepsin enzyme, flours treated by hexane, starch, flours treated Fig. 2. The index digestibility of millet flours at different by pepsin enzyme and hexane solution, raw starch), the treatments.

Table 1. Chemical composition of sorghum and millet flours. Composition, % Sorghum flour Millet flour humidity 9.81 9.37 ash 1.61 2.04 Solubles sugar 2.48 3.63 insoluble fibres 5.52 6.64 protein 12.27 11.17 lipid 4.37 6.64 total starch 66.37 65.21

65 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

Table 2. The in vitro starch digestibility of sorghum flours at different treatments. Sample Starch initial weight, Starch digested weight, mg (a) Index digestibility, % mg (b) SR1 10 3.64±0.10 36.4 SR 2 10 5.54±0.47 55.4 SR 3 10 3.66±0.10 36.6 SR 4 10 8.76±0.29 87.6 SR 5 10 8.95±0.40 89.5

Table 3. The in vitro starch digestibility of millet flours at different treatments.

results are presented in Table 2. Fig. 1 demonstrates in millet (the sample 5, starch isolation) giving 89.7 % vitro starch digestibility of sorghum starches at differ- extent of digestibility for millet and 89.5 % for sorghum. ent treatments. Table 3 and Fig. 2 present in vitro starch The presence of protein in sorghum and millet flours digestibility of millet starches at different treatment. have an impact on the starch digestibility by The digestibility of sorghum flour is 36.4 % and amyloglucosidase enzyme at different pretreatments, 36.8 % for millet as compared with the results of Snow and where the digestibility was more increased after re- O’Dea, the starch hydrolysed of barley flour is 15.06 %, moval of some proteins by pepsin (sample 2), 56.6 % rolled wheat 5.05 % and rye flour 18.63 % [21]. After for millet and 55.4 % for sorghum, respectively. The low hydrolysis by amylase and glucoamylase at 50°C for 30 starch digestibility of cooked sorghum flours was stud- min the digestibility of wheat flour by glucoamylase ied by Zhang and Hamaker [8]. enzyme is 15 % [18]. This comparison indicates that our sorghum and millet flour are highly hydrolysed in Effect of the protein and the lipid on the starch di- vitro by glucoamylase. gestibility The results indicate that the amyloglucosidase is The influence of the protein and the lipid on the able to hydrolyse rapidly the starch of sorghum and starch digestibility is illustrated in Table 4. According

Table 4. The impact of protein and lipid on sorghum and millet samples.

Impact of protein (%) Impact of lipid (%) (a) (b) sorghum 52.9 33.2 millet 53.3 34.1

(a) impact of protein is the relative digestibility of sample (5) minus the index digestibility of sample (3); (b) impact of protein is the relative digestibility of sample (5) minus the index digestibility of sample (2).

66 Rachid Souilah, Badreddine Belhadi, Nadia Boudries, Djaafar Djabali, Boubekeur Nadjemi

to the impact of protein and lipid of the sorghum and starch, Food Chemistry, 65, 1999, 417-425. the millet samples, we notice the that the digestibility is 5. L. W. Rooney, R. L Pflugfelder, Factors affecting increasing after the removal of some protein and lipid starch digestibility with special emphasis on sorghum and the impact of protein is 53 % and of lipid is 34 %. and corn, J. Anim. Sci., 63, 1986, 1607-1623. These findings indicate that the protein of sorghum and 6. K. E. Bach Knudsen, L. Munck, B. O. Eggum, Effect millet plays an important role in slowing the starch di- of cooking, pH and polyphenol level on carbohy- gestion. drate composition and nutrition quality of sorghum Generally, we can say that the physico-chemical (Sorghum bicolor (L.) Moench) food, ugali, Br. J. treatment of sorghum and millet flours plays a role in Nutr., 59, 1988, 31-47. the industrial processes to convert the starch into 7. K. E. B. Knudsen, A. W. Kirleis, B. O Eggum, L. dextrines on glucose sugar and influences the hydrolysis Munck, Carbohydrate composition and nutritional of starch by glucoamylase enzyme. quality for rats of sorghum to prepared from decor- ticated white and whole grain red flour, The journal CONCLUSIONS of nutrition,118, 1988, 588-597. 8. G. Zhang, B. R. Hamaker, Low -Amylase Starch Di- Based on the results, we deduced that the pre- gestibility of Cooked Sorghum Flours and the Effect treatments effectuated on sorghum and millet flours has of Protein, Cereal Chem., 75, 1988, 1, 710-713. a clear impact on the digestibility. Also, the presence of 9. A. O. Elkhalifa, B. Schiffler, R. Bernhard, Effect of proteins had a greater impact in lowering the starch fermentation on the starch digestibility, resistant digestibility than lipids. Their presence reduces the starch and some physicochemical properties of sor- digestibility as a resistant wall and hindered the passage ghum flour, Nahrung/Food, 48, 2004, 2, 91-94. of the enzyme solution to starch. They increase the 10. A. Urooj, Sh. Puttraj, Digestibility index and factors af- susceptibility in the case of isolated starch (sample 5). fecting rate of starch digestion in vitro in conventional Besides the starch in sorghum and millet has a great food preparation, Nahrung., 43, 1, 1999, 42 - 47. digestibility by glucoamylase (>89 %). 11. B. Godon, W. Loisel, Guide Pratique D’Analyse Dans Les Industries de céréales. Technique et Docu- Acknowledgements mentation,1984. We express our gratitude to the laboratory of 12. D.A.T. Southgate, Determination of food carbohy- bioactive products and biomass valorization research. drate, applied science publishers LTD., London,1976. (ENS Kouba, Alger). 13. I. Goni, A. Garcia-Alonso, F. Saura-Calixto, A starch hydrolysis procedure to estimate glycemic index, REFERENCES Nutrition Research., 17, 3, 1997, 427-437. 14. E. T. Mertz, M. M. Hassen, C. Cairns-Whittern, A. 1. P.S. Belton, J.R.N. Taylor, Sorghum and millets: pro- W. Kirleis, L. Tu, J. D. Axtell, Pepsin Digestibility tein source for Africa Trends in food science and of Proteins in Sorghum and Other Major Cereals, technology, 15, 2004, 94-98. Proc, Natl, Acad, Sci., 81, 1984, 1-2. 2. Y. Granfeldt, I. Brorck, A. Drews, An in vitro pro- 15. J. Toivo, A. M. Lampi., S. Aalto, V. Piironen, Fac- cedure based on chewing to predict metabolic re- tors affecting sample preparation in the gas chro- sponse to starch in cereal and legume products, matographic Determination of plant sterols in whole European journal of clinical nutrition., 46, 9, 1992, wheat flour, Food Chemistry, 68, 2000, 239-245. 649-660. 16. T. Beta, H. Corke, L. Rooney, J. R. N. Taylor, Starch 3. F. Bornet, Technological treatments of cereals. Re- properties as affected by sorghum grain chemistry, percussions on the physiological properties of starch, Journal of the Science of Food and Agriculture, Carbohydrate Polymers., 21,1993, 2-3, 195-203. 81, 2001, 2, 245-251. 4. R. Cui, C. G. Oates, The effect of amylose-lipid com- 17. T. Beta, H. Corke, Genetic and environmental varia- plex formation on enzyme susceptibility of sago tion in sorghum starch properties, Journal of cereal

67 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

science, 34, 3, 2001, 261-268. 36, 1985, 1019–1027. 18. T. Noda, S. Takigawa, C. Matsuura-Endoa, K. Saitoa, K. 20. I. Barikmo, F. Ouattara, A. Oshaug, Protein, Carbo- Takata, T. Tabikia, H. Arachichige, M. Wickramasingheb, hydrate and Fibre in cereals from Mali-How to fit H. Yamauchia, The physicochemical properties of the results in a food composition table and data- partially digested starch from sprouted wheat grain, base, Journal of food composition and analysis., 17, Carbohydrate Polymers, 56, 2004, 271–277. 2004, 291-300. 19. J.J. Karkalas, An improved enzymatic method for 21. P. Snow, K. O’Dea, Factors affecting the rate of the determination of native and modified starch, hydrolysis of in food, the American journal of Clinical Journal of the Science of Food and Agriculture, Nutrition., 34, 1981, 2721-2727.

68 Journal of the UniversityReyad of A.Chemical Al Dwairi, Technology Aiman E. and Al-Rawajfeh Metallurgy, 47, 1, 2012, 69-76

REMOVAL OF COBALT AND NICKEL FROM WASTEWATER BY USING JORDAN LOW-COST ZEOLITE AND BENTONITE

Reyad A. Al Dwairi1, Aiman E. Al-Rawajfeh2

Tafila Technical University, P.O. Box 179, Received 25 February 2011 Tafila 66110, Jordan. Accepted 10 january 2012 1 Natural Resources and Chemical Engineering Department E-mail: [email protected] 2 Energy and Oil Shale Researche Center (EOSRC) E-mail: [email protected]

ABSTRACT

In this work, a mixture of natural zeolitic tuff (Z) and bentonite (B) from Jordan were used to treat wastewater from cobalt and nickel. The results showed that bentonite type performed better efficiency in heavy metal removal than the zeolite type, while the 50/50 Z/B mixture type shows the highest percentage removal. The high values of correlation factor indicate that the adsorption onto Z/B mixture is more accurately described by Freundlich isotherm. The negative value of ÄGo confirms the feasibility of the process and the spontaneous nature of adsorption. The removal process was well described with the second-order reaction kinetics. Keywords: heavy metals, zeolite, bentonite, adsorption isotherms, wastewater treatment.

INTRODUCTION and kaolinite as clay minerals, while diatomite, evaporates (halite, gypsum with bassanite and anhydrite), Zeolites are hydrated aluminosilicate minerals quartz, dolomite, feldspar and calcite as non-clay (natural or manufactured) usually with a 3-D structure minerals [2]. Bentonite is a known adsorbent that has based on polyhedra [SiO ]4- – [AlO ]5- networks. They been used in the management of paraquat poisonings. 4 4 have a unique structure and characteristics that make Al-Anber [3] used Jordanian bentonite (NB) and them to adsorb effectively a wide range of environmental quartz (NQ) for the removal of high-level Fe3+ from pollutants. Their deep and wide pore openings are just aqueous solution. Bourassa et. al. [4] used zeolite and one of few characteristics which enable to remove bentonite as a micro particle system in the clarification various contaminants. Another characteristic of zeolite of drinking water accompanied with dissolved air is their large surface area (20-50 m2/g by natural species). flotation. Park et. al. [5] performed a study of activated- Also natural zeolites are characterized by availability sludge and a comparison between zeolite and activated and low mining cost, bulk density, high resistance to carbon carriers on the nitrification process. In Drag’s et alteration [1]. There are two types of bentonite: (1) the al. [6] study, zeolite-carbon adsorbents were used to treat sodium, high-swelling type, derived from volcanic ash wastewaters from carbonaceous deposits. Princz [7] that was deposited in marine environments; and (2) the studied the improvement of biological degradability of calcium, low-swelling type, which evolved from volcanic wastewaters using activated zeolites. In the course of this ash deposited in freshwater environments. The treatment process, zeolite was added to the influent water investigation revealed the presence of illite/smectite of or directly into the aeration basin of an activated sludge different expendabilities, montmorillonite, palygorskite system. The zeolite increased ammonium removal

69 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012 efficiency and the decomposition rate of the organic applications in wastewater treatment; he concluded that matters as well as the settling characteristics of the Al Ataita phillipsitic tuff shows good removal percent- activated sludge (sludge-volume index, floc size). ages of ammonium from municipal wastewater. Al Clinoptilolite zeolite was recommended for removal of Dwairi [20] used Uniza zeolitic tuff for industrial waste- Cs+, Ni2+, Sr2+ and Ba2+ from radioactive or municipal water treatment; he found that Uniza zeolitic tuff shows wastewater. Tian et. al. [8] proposed the removal of good removal percentages of Cd and Mn from indus- organic compounds and ammonia from municipal trial wastewater. wastewater by using biological filtration with zeolite The purpose of this work is to study the Jordanian medium to elongate the life of the zeolite examined. natural zeolites (Z), bentonite (B) and their mixture (Z/ Ellis et. al. [9] tried to remove organics from retort waters B) as potential adsorbent materials for Co and Ni derived from Stuart Oil Shale using the high silica removal from wastewater. zeolites. Removal of organics increased with the decreasing polarity of the adsorbate and was improved EXPERIMENTAL when free ammonia was firstly removed by air stripping. Raw Materials Zeolites were more effective than activated carbon, and This study focuses on testing the Jordanian natu- were potentially useful for pretreatment of retort water ral bentonite and zeolite raw materials and their mix- before biological treatment. tures as an adsorbent media. For this purpose, crushed Jordan zeolite was evaluated as an adsorbent natural zeolite (Phillipsite) is amended with commer- material in wastewater treatment especially the N-E type cial powdered bentonite to yield a soil mixture low in (arytane) [10-12]. Attilli [13] has evaluated the Jordanian permeability and high in ion-exchange capacity. Con- phillipsitic tuff in water softening. Laboratory and pilot centrations of certain heavy metals in the effluent fluid plant scale removal of ammonium and toxic heavy metal percolated through the 50/50 % bentonite/zeolite mix- ions were studied. Ibrahim et. al. [14] and Ed-Deen [15] ture are compared with that of initial concentrations. have evaluated zeolites from Tell Rimah for treatment The raw materials in this study were analyzed by of industrial wastewater from Electroplating and Battery the Scanning electron microscope (SEM) and X-ray Factory effluents, Tell Rimah zeolitic tuff showed high difractometer. Zeolites are of phillipsitic type [17], while selectivity for the removal of Cu, Cr, Ni and Zn and bentonite was studied by Nawasreh [21]. The could be used for removing Pb and Fe. Ibrahim and characteristics of zeolite and bentonite samples used in Akashah [16] found that faujasite tuff from Jordan has this study are available in the literature [22-23]. a good selectivity for Lead from wastewater. Al Dwairi [17] studied the characterization of Sample Preparation and Experimental Method Jordanian zeolitic tuff and its potential use in wastewater Two Bulk samples of natural raw geomaterials treatment, he studied the mineralogy and geochemistry (zeolite and bentonite) were mixed thoroughly after of all zeolitic tuff localities in Jordan and evaluated sieving. The bulk sample was sieved by different size of Jordanian natural zeolitic tuff obtained from Hannoun sieves and the obtained fractions. The size (1-0.3 mm) (HNZ) and Mukawir (MNZ) volcanoes for removing of was used in the experiments which are characterized by selected heavy metals (Zn and Pb) and total organic high zeolite content, high attrition resistance, backed carbon (TOC) from domestic wastewater. The results bed densities and good permeability [17]. show that the main dominated zeolites minerals in For determining the amount of heavy metal re- Jordan are chabazite, phillipsite, faujasite and analcime. moval by different adsorbents, batch experiments were The results of wastewater treatment show that HNZ and used; it consists of the addition of a volume (50 ml) of MNZ is stubble for domestic wastewater treatment. wastewater to adsorbent material (10g) for a certain time Ibrahim and Jbara [18] have used natural phillipsite- (1-12 hours), with continuous stirring, until a condi- faujasite tuff from Jordan to remove paraquat from tion of equilibrium is established in the exchange reac- synthetic wastewater. Al Dwairi [19] evaluated the tion. The filtrates were analyzed for Co and Ni by atomic Jordanian phillipsitic tuff from Al Ataita volcano for absorption spectroscopy.

70 Reyad A. Al Dwairi, Aiman E. Al-Rawajfeh

RESULTS AND DISCUSSION mixtures are calculated and drawn for cobalt in Fig.1, and for nickel in Fig.2. Metal ions removal and adsorption isotherms The uptake of cobalt or nickel ions was calcu- The amount of cobalt or nickel removal by during lated from the mass balance, which was stated as the the series of batch investigations was determined using amount of solute adsorbed onto the solid. It equals the the following equation expressed as: amount of solute removed from the solution. The adsorbed amount (Q ) per unit absorbent mass (mg/g) e − was calculated as follows: C0 Ce Re moval[%] = ×100 (1) − C (C0 Ce )V 0 Qe = (2) m where C and C are the initial and equilibrium concen- where C is the initial heavy metal concentration, C is the 0 e 0 e tration (ppm) of metal ions in solution, respectively. concentration of heavy metal at equilibrium (mg l-1), m is The percentage removals by zeolite, bentonite and their the clay mass (mg) and V is the solution volume (l).

100 99 98 97 96 95

2+

% Removal 94 % Removal of Co after 12 h 93 Z 92 B 91 Z/B 90 0 1020304050 Initial Conc. [ppm]

Fig. 1. Percentage removal of Co2+ from different initial concentrations after 12 h on zeolite, bentonite and their mixtures.

100 90 80 70 60 2+ 50 % Removal of Ni after 12 h

% Removal 40 Z 30 B Z/B 20 10 0 0 1020304050 Initial Conc. [ppm]

Fig. 2. Percentage removal of Ni2+ from different initial concentrations after 12 h on zeolite, bentonite and their mixtures.

71 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

30 4.5 4.0 Co Co 25 Ni 3.5 Ni

20 3.0 e e 2.5 /Q

e 15 C ln Q ln 2.0

10 1.5

5 1.0 0.5 0 0.0 0 102030405060 012345 Ce ln Ce Fig. 3. Langmuir adsorption model for interactions of Co2+ Fig. 4. Freundlich adsorption model for interactions of Co2+ and Ni2+ on Z/B mixture. and Ni2+ on Z/B mixture.

The Langmuir isotherm is based on the theoreti- and 0.85 for cobalt and nickel metal ions, respectively. cal principle that only a single adsorption layer exists For the favorable adsorption R value must take place 0 L on an adsorbent and it represents the equilibrium dis- < R <1. In our results, R values confirmed that L L tribution of metal ions between the solid and liquid Langmuir isotherm was favorable for the adsorption of phases. The basic assumption of the Langmuir both cobalt and nickel Z/B adsorbent. adsorption process is the formation of a monolayer of Freundlich isotherm is used for modeling the ad- adsorbate on the outer surface of the adsorbent and after sorption on heterogeneous surfaces. This isotherm can that no further adsorption takes place. The attraction be explained by the linear form [25]: between molecules decreases as getting further from the 1 logQ = log K + logC adsorption surface [24]. The Langmuir-type isotherms e f n e (5) remain to be the most widely used for practical where K is the Freundlich constant (mg/g) and 1/n is f applications. Langmuir isotherm can be defined the adsorption intensity. according to the following linear form: Fig. 4 shows the dependence of lnQ from lnC . e e The correlation factor is almost 1 for both cobalt and nickel. The Freundlich adsorption isotherm is an Ce 1 Ce = + (3) indicator of the extent of heterogeneity of the adsorbent Qe kV V m m surface. This Freundlich type behavior is indicative of where V is the monolayer capacity and k is the equilib- the surface heterogeneity of the adsorbents, i.e. the m rium constant. adsorptive sites (surface of Z, B and Z/B) are made up The dependence of C / Q from C , obtained by of small heterogeneous adsorption patches that are e e e using experimental results, is shown in Fig. 3. The cor- homogeneous in themselves. The activation of adsorption relation factors are 0.998 and 0.967 for cobalt and nickel, site takes place, leading to increased adsorption probably respectively. Langmuir isotherm constants are shown in through the surface exchange mechanism. Freundlich Table 2. isotherm constants are shown in Table 1. Langmuir isotherm determines the adsorption Langmuir and Freundlich isotherms are favorable or unfavorable. To determine the characteristic insufficient to explain the physical and chemical behavior of the adsorption, dimensionless equilibrium characteristics of adsorption. Dubinin–Radushkevich parameter is used. The equation is given as: (D–R) isotherm is commonly used to describe the 1 sorption isotherms of single solute systems [26]. The R = (4) D–R isotherm, apart from being analogue of Langmuir L 1+ kC 0 isotherm, is more general than Langmuir isotherm as it where k is the Langmuir constant and C is the highest rejects the homogeneous surface or constant adsorption 0 metal ion concentration. The R values were found 0.64 potential. The D–R isotherm is expressed as: L 72 Reyad A. Al Dwairi, Aiman E. Al-Rawajfeh

Table 1. Langmuir, Freundlich and Dubinin–Radushkevich (D–R) isotherms parameters of cobalt and nickel removal on Z/B mixture.

Metal Langmuir isotherm Freundlich isotherm D-R isotherm

2 2 2 Vm k R Kf n R K' V'm E R mgg-1 mol2 kJ-2 mg g-1 kJmol-1 Co2+ 2.73 0.04 0.998 0.87 0.98 0.999 -0.52 30.00 0.98 0.939 Ni2+ 33.00 0.01 0.967 0.54 0.92 0.999 -0.21 4.06 1.54 0.954

= ′ − ′ε 2 (6) nickel are calculated as 0.98 and 1.54 (kJ/mol). The lnQe lnVm K high values of R2 (> 95%) indicates that the adsorption of cobalt and nickel onto Z, B, and Z/B mixture is well where V ′ is the D–R adsorption capacity (mg/g),− K ′ε is m described by the three isotherms, but more accurately the constant related with adsorption energy (mol2/kJ2), (R2 values are almost 100%) with Freundlich isotherm. and å is the Polanyi potential. According to the Eq. (6), the Polanyi potential (å) can be given as: Kinetics of adsorption 1 The study of sorption kinetics describes the up- ε = RT ′ln(1+ ) (7) Ce take rate of cobalt and nickel ions, and evidently this rate controls the residence time of these ions at the solid -1 -1 ′ where R is the gas constant (kJK mol ) and is the liquid interface. Consequently it is important to establish temperature (K). The main energy of adsorption (E) is the time dependency of such systems for various calculated by using the following formula: pollutant removal processes. Therefore, the required 1 contact time for the sorption to be completed is E = (8) − 2K ′ important to give insight into a sorption process. This also provides information on the minimum time required where E gives information about the physical and for considerable adsorption to take place and the possible chemical features of adsorption. diffusion control mechanism between the adsorbed ion The D–R isotherm is applied to the data ob- as it moves from the bulk solution towards the adsorbent tained from the experimental studies. A plot of lnQ e surface. At the beginning stage of the adsorption process, 2 against å yields a straight line and it is given in Fig. 5 the removal rate of the ions is higher. The faster beginning 2 (R values are about 94% and 95%, for cobalt and nickel, rate may be due to the availability of the uncovered respectively). The adsorption capacities ( ′ ), adsorp- Vm surface area of the Z and B initially. The adsorption ′ tion energy constants ( K ε) and the main adsorption en- kinetics depends on: (i) the surface area of the adsorbent, ergies (E) are shown in Table 1. (ii) the nature and concentration of the surface groups All of the isotherm model parameters for the ad- (active sites), which are responsible for interaction with sorption of cobalt and nickel are provided in Table 1. targeted ions [3]. The kinetics of cobalt and nickel ions Langmuir isotherm calculated adsorption capacities for sorption on the Z/B adsorbent mixture was analyzed cobalt and nickel are 2.73 and 33.00 mg/g, respectively. using two kinetic models, these include: the pseudo- The negative value of ÄGo confirms the feasibility of first order and pseudo-second order. the process and the spontaneous nature of adsorption The adsorption rate constant proposed by Ho [27] with a high preference for Z/B adsorbent to remove using first order reaction kinetic is shown below: nickel metal ions (ÄGo = -11.74 kJ/mol) in compari- dQ o t = − son to cobalt metal ions (ÄG = -8.99 kJ/mol). In k1 (Qe Qt ) (9) Freundlich isotherm adsorption intensities are found dt to be 0.98 and 0.92 for cobalt and nickel metals. By where k is the adsorption rate constant for the first 1 using D–R isotherm, adsorption energies for cobalt and order adsorption, and Q is the amount of heavy metal t

73 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

5 1 4.5 = k t + C Co − 2 2 (12) 4 Ni Qe Qt 3.5 where C is the integration constant of the second order 3 2 e 2.5 reaction kinetic. ln Q 2 For both metals, it is seen that the curves in the 1.5 − 1 plots of 1/(Qe Qt ) against time, are linear. Second 0.5 order reaction kinetics for cobalt and nickel adsorption 0 onto Z/B is shown in Fig. 7. 0 0.2 0.42 0.6 0.8 1 ε 2 Rate constants were calculated from the slopes of the curves (Table 2). The degree of goodness of the Fig. 5. Dubinin–Radushkevich adsorption model for linear plot of these kinetic models can be judged from interactions of Co2+ and Ni2+ on Z/B mixture. the value of the determination coefficient of the plot, which can also be regarded as a criterion in the

2.0 determination of the adequacy of kinetic model. From 1.8 the determination coefficient values above, adsorption 1.6 1.4 of cobalt and nickel ions on Z/B adsorbent is regarded

) 2 t 1.2 as pseudo-second order (R > 99 %) rather than pseudo-

-Q Co e 1.0 Ni first order (R2 values are about 85 % and 97 % for

ln (Q ln 0.8 0.6 cobalt and nickel, respectively). 0.4 0.2 CONCLUSIONS 0.0 0 50 100 150 200 250 300 Time [min] In this study, the removal of Co2+ and Ni2+ was investigated using zeolite, bentonite and their 50/50 Fig. 6. Pseudo-first order reaction kinetics for the adsorption mixture. The experimental values were evaluated of cobalt and nickel on Z/B. according to the Langmuir, Freundlich and D–R isotherms that are generally used to describe the adsorption processes. It was found that all of isotherm adsorbed at time t (mg/g). The integration of the Eq. (9) models fit very well. The correlation coefficient for co- gives the following expression: balt and nickel Freundlich isotherm adsorption was the − = − + highest. The Langmuir isotherm adsorption capacities ln(Qe Qt ) k1t C1 (10) where C is the integration constant for first order reac- 1 tion kinetic. 25 For both metals, it is seen that the curves in the Co 20 Ni − plots of ln(Qe Qt ) against time, are linear. First order ) reaction kinetics for cobalt and nickel adsorption onto t 15 -Q Z/B is shown in Fig. 6. e

1/(Q 10 Adsorption data was also evaluated according to the Pseudo-second order reaction kinetic proposed by 5 Ho and McKay [28]: 0 0 50 100 150 200 250 300 dQt 2 = k (Q − Q ) (11) Time [min] dt 2 e t where k is the second order reaction constant. If Eq. 2 Fig. 7. Pseudo-second order reaction kinetics for the (11) is integrated, the following expression is obtained: adsorption of cobalt and nickel on Z/B.

74 Reyad A. Al Dwairi, Aiman E. Al-Rawajfeh

Table 2. Pseudo-first and pseudo-second order reaction kinetics of cobalt and nickel removal on Z/B mixture. Metal Pseudo-first order Pseudo-second order -1 2 -1 -1 2 k1 ,1 min R k2 g mg min R Co2+ 0.0008 0.850 0.0124 0.998 Ni2+ 0.0068 0.970 0.0033 0.999 for cobalt and nickel were 2.73 and 33.00 mg/g, respec- aqueous solution using natural inorganic materials: tively. The negative value of ÄGo confirms the feasibil- bentonite (NB) and quartz (NQ), Desalination, 2010, ity of the process and the spontaneous nature of adsorp- 250, 885–891. tion with a high preference for Z/B adsorbent to remove 4. C. Bourassa, M. Sain, C. C. Daneault, Chem. Technol., nickel metal ions more than cobalt metal ions. The 36, 2003,193-202. Freundlich isotherm adsorption intensities for cobalt 5. S. J. Park, J. W. Oh, T. Yoon, Process Biochem., 39, and nickel were 0.98 and 0.92, respectively. Moreover, 2003, 211-219. the D–R isotherm adsorption energies were calculated 6. E. B. Drag, Pr. Nauk, Chem. Technol. Nafty Wegla to state the physical and chemical characteristics of Politech. Wroclaw, 57, 2000, 209-215. adsorption. The magnitudes of E for cobalt and nickel 7. Princz, A Study Supported by NATO’s Science for adsorption were 0.98 and 1.54 kJ/mol, respectively. Peace Programme, NATO Project # SFP 972494, These low values of adsorption energy show that the 1999. adsorption has a physical nature. The removal of Co2+ 8. W. H. Tian, X. H. Wen, Y. J. Qian, Environ. Sci., 16, and Ni2+ from wastewater using Z/B mixture was well 2004, 90-93. described with the second-order reaction kinetics. The 9. J. Ellis,J. Korth, L. Peng, Fuel, 74, 2001, 860-864. correlation coefficients were higher for the second-or- 10. I. Dwairi, A chemical study of the palagonitic tuffs der reaction kinetics. As a result of this study, it may be of the Aritain area of Jordan, with special reference concluded that the low-cost, abundant and locally avail- to nature, origin and industrial potential of the able zeolite, bentonite and their mixtures, can be used associated zeolite deposits, PhD thesis, Hull Univ., for elimination of cobalt and nickel heavy metals from UK, 1987. wastewater. 11. I. Dwairi, Evaluation of Jordanian phillipsite tuff in removal of ammonia from wastewater: Experimental Acknowledgements study. Al- Belga Journal, 1991, 153-66. Abdul Hameed Shoman Foundation is greatly 12. I. Dwairi, Jordanian zeolites: Evolution for possible acknowledged for funding this research project. industrial application of natural Aritain phillipsite tuffs, Dirasat, 18, 1992, 23-44. REFERENCES 13. Sh. Attili, Evaluation of Jordanian phillipsite tuff, with special regards to its application in water soft- 1. B. Mercer, L. Ames, Zeolite ion exchange in radioac- ening, removal of ammonium and toxic heavy metal tive and municipal wastewater treatment: In sand, L. ions, M.Sc. thesis, Yarmouk University, Irbid, Jor- B. and Mumpton, F. A. Natural zeolites: Occurrence, dan, 1992. Properties and Uses. Pergamon Press, 1978, p. 451- 14. M. K. Nawasreh, F. D. Pooley, Mineralogical distribu- 462. tion and industrial evaluation of Azraq clay minerals 2. J. Alali, A. Abu Salah, Exploration for Bentonite and in the North Badia Region, Jordan, Presented in the other Minerals in Azraq Depression. Natural Re- Applied Geology Conference, Amman, Jordan, 1998. sources Authority. Unpublished Internal Report, Jor- 15. T. Ed-Deen, Zeolites from Tell Rimah and its use dan, Amman, 1993. for industrial wastewater treatment. unpublished M. 3. M. A. Al-Anber, Removal of high-level Fe3+ from Sc. thesis, University of Jordan, 1998.

75 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

16. K. M.Ibrahim, T. Akashah, Lead removal from Environmental Geology, 41, 2002, 547-551. wastewater using faujasite tuff, Environmental Ge- 22. http://www.nra.gov.jo/images/stories/pdf_files/ ology, 46, 2004, 865-870. Bentonite.pdf (Visited 17/1/2012). 17. R.A. Al Dwairi, Characterization of the Jordanian 23. http://www.nra.gov.jo/images/stories/pdf_files/ zeolitic tuff and its potential use in Khirbet es Samra Bentonite.pdf (Visited 17/1/2012). wastewater treatment plant. Ph.D. thesis, The Uni- 24. N. Unlu, M. Ersoz, Adsorption characteristics of versity of Jordan, Amman, Jordan, 2007. heavy metal ions onto a low cost biopolymeric sor- 18. K.M. Ibrahim, H.A. Jbara, Removal of paraquat from bent from aqueous solutions, J. Hazard. Mater., synthetic wastewater using phillipsite-faujasite tuff B136, 2006, 272–280. from Jordan, J.Hazard. Mater., 163, 2009, 82-86 25. J. Zeldowitsch, Über den Mechanismus der 19. R.A. Al Dwairi, The use of Expendable Local Zeo- katalytischen Oxydation von CO an MnO , Acta 2 lite Deposits for NH Removal in Municipal Waste- Physicochim., URSS 1, 1934, 364–449. 4 water, Jordan Journal of Civil Engineering, 3, 2009, 26. S. Veli, B. Alyuz, Adsorption of copper and zinc 3. from aqueous solutions by using natural clay, J. 20. R.A. Al Dwairi, Mn+2 and Cd+2 removal from in- Hazard. Mater., 149, 2007, 226–233. dustrial wastewater using phillipsitic tuff from Jabal 27. Y.S. Ho, Citation review of Lagergreen kinetic rate Uniza, Southern Jordan, Jordan Journal of Civil equation on adsorption reaction, Scientometrics, 59, Engineering, 4, 2010, 1. 2004, 171–177. 21. K. M . Ibrahim, T. Nasser Ed-Deen, H. Khoury, Use 28. Y.S. Ho, G. McKay, Sorption of dye from aqueous of natural chabazite-phillipsite tuff in wastewater solution by peat, J. Chem. Eng., 70, 1998, 115– treatment from electroplating factories in Jordan, 124.

76 Journal of the University of ChemicalA. TechnologyAtanassov and Metallurgy, 47, 1, 2012, 77-82

ADVANCED SOFTWARE ARCHITECTURE OF AN AUTOMATIC VEHICLE NUMBER PLATE RECOGNITION SYSTEM

A. Atanassov

University of Chemical Technology and Metallurgy Received 09 November 2011 8 Kl. Ohridski, 1756 Sofia, Bulgaria Accepted 10 January 2012 E-mail: [email protected]

ABSTRACT

The paper presents the selection and development of a flexible software architecture intended to an Automatic Number Plate Recognition (ANPR) System. The capabilities of the available on the market competitors’ ANPR systems were analyzed and the requirements to the new generation of ANPR system were recognized and/or formulated. Different software architectures, matching the requirements, were analyzed and the WEB–based one was selected. The appropriate software components corresponding to the requirements were identified.. The components were developed using the process-oriented approach based on the Theory of the Communicating Sequential Processes (CSP). All ANPR components were decomposed to the CSP processes exchanging specific messages via channels. Finally, these processes were mapped to the operating system’s processes and threads communicating in parallel. As a result an advanced reconfigurable ANPR System minimizing the execution time was provided to the customers. Keywords: ANPR system, WEB based UI, traffic control software, CSP.

INTRODUCTION • recognition or not of the number plate via OCR [3]; • verification of the found plate number using plates The automatic number plate recognition (ANPR) data in DB; systems become of great importance for on-line traffic • and visualizing the found number (and car image, im- control. There are different types of ANPR systems age stream) to the operator. deployed in special scopes: traffic control, traffic • In some cases sending some data or signals to the digi- violation, control of parking lots, highway lots, control tal outputs or reading additional data from the inputs. of the access to cities centers, etc. Those systems, in Most of the existing ANPR systems [4] are with general, are based on the Charge-Coupled Device (CCD) specific purposes corresponding to one or more of infrared or color cameras, radars (optional), some digital mentioned above scopes. In most of them the camera inputs and outputs, Optical Character Recognition hardware and the computer(s) running the software (OCR) software for number plate recognition, Data Bases (OCR, GUI and DB) are placed on different places and (DB) and Graphical User Interface (GUI) for interaction the connection between them is based on coaxial or to the operators. Usually, the process (Fig. 1) of number optical cables. A great majority of systems are still using plate recognition [1, 2] includes: this architecture. This is no robust architecture because • getting image stream from the camera and sending it if the PC controlling the camera fails the whole ANPR to the operator PC; system fails. It requires complex installation and start- • converting the image stream to the sequence of im- up. It is necessary to install video and control wire for ages in appropriate image format (TIF, GIF, JPG, BMP) each one of the cameras, furthermore, it is necessary to used by the OCR software. It is done by SW for image provide power supply cable to the cameras. If the dis- handling (IH); tance between the lanes is too long, the signal of the

77 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

munication with the client application. It is modular ar- chitecture and if an ANPR system, controlling one road lane, does not work, its fall does not affect to the other lanes. Installation and start-up is easier. It is only neces- sary to provide 220v, Ethernet network or serial com- munication to each embedded system. The installation and the maintenance of the ANPR system are reduced. Current paper presents the new flexible architec- ture of an embedded ANPR system. The system is built by set of Operating System (OS) processes related to the Fig. 1. Hardware and software components of ANPR system. required ANPR functionalities. The proposed architec- ture is developed on the base of theory of the Communi- cating Sequential Processes (CSP). The ANPR system provides WEB-based GUI component for remote (Internet) operative control, adjustment and reconfiguration of all HW and SW components forming the ANPR system.

SOFTWARE REQUIREMENTS TO ANPR SYSTEMS

After intensive investigations and analysis of the leading ANPR systems vendors Tattile, Vitronic [6], Leutron Vision [7], Hi-Tech Solutions [8], etc. the general and specific functional requirements valid to all contemporary ANPR systems were defined. General requirements: • Online capturing of the images of the vehicles’ Fig. 2. All-in-one ANPR Systems. number plates. • Online optical character recognition (OCR) of cameras will not arrive with clearness to the PC. This the captured vehicles’ number plates. architecture is very expensive because of the wiring and • Providing an operator control of the observed the time of installation is multiplied, as well the by the ANPR system object, including: maintenance of the system is multiplied. • Supervising the ANPR live video-stream of the Other ANPR systems are based on cameras with controlled road, lane, etc. software that implements some IH and OCR procedures. • Visualization of the found number plates Last generation of ANPR systems (Fig. 2) are embedded • Support of databases with information of the systems that incorporate in one body the camera (infra- vehicle plate numbers (Black and White lists) and red, color or both), the illumination and the PC hard- providing the search in DB in order to find if there is a ware with external interfaces, power supply, the operat- hit (match) of the found plate number in the White/ ing system, IH, OCR and communication SW. They are Black list. known as all-in-one ANPR Systems. These systems are • Collecting and sending of plain or encrypted capable immediately to inform other external systems or information (evidential records (ER)) of each plate interested organization for vehicles data. number to the specific database servers via internet or The advantages of this architecture are that it is GPRS. simpler [5] one and all the necessary elements for the • Configuration of the ANPR camera parameters, ANPR system are integrated in the same housing. The as brightness, gain, frames per second, shutter, position, equipment may be connected by Ethernet or serial com- angles, etc. Tuning of OCR parameters.

78 A. Atanassov

• Restart of the ANPR application or Reboot of the operating system. • Authentication of logging ANPR users. • Reporting the status of entire ANPR systems (its specific hardware devices or software components) to the operator. Specific requirements: • Configuration of the interfaces to the external systems (DB servers, GPRS, etc.) • Adjustment of serial lines (COM ports) in order to control parking or highway lots Fig. 3. CSP Processes forming ANPR System. • Download and upload of configuration files or software updates. - Image Handler (IH) Process is responsible • Support of configurable rules and actions related for getting raw live images from the IR or color cameras to found plate numbers, meaning how and where the and transforming them into GIF, TIF or JPEG formats ERs should be saved or which external system should used by the OCR process. The IH also redirects non- be triggered. transformed images to ANPR Manager Process. • Collecting of statistical data (recognized/not - The OCR process produces the results: – finds recognized numbers per hour, day). plate numbers of the observed vehicles. These results • Collecting of logging and tracing data. are directed to ANPR Manager Process for further • Support of recording of test decks with live- handling. video images used for further adjustment of the OCR - The ANPR Manager Process (AM) transforms algorithm. the live images to the image formats used by the Web Server Process and sends them to it. It gets the found IDENTIFYING CSP PROCESSES number plates data and produces the ER images which are sent to the Web Server and/or to the remote DB The CSP theory [9] deals with processes that are Server. AM is also responsible for checking whether communicating events or messages over channels. The the found number plate matches the ones given into the processes interact with each other sequentially, in parallel Black or White verification lists, or if the number or alternatively. Some processes can be decomposed to matches the rules and actions formed via Web UI. It other (sub) processes and channels and etc. The also sends the status data from Status Manager Process decomposition approach leads to refining of the inter- to the Web Server and DB Server. AM is responsible processes relations and communications. Each process for I/O ports control. implements its own internal algorithm which is not - The Status Manager Process collects hardware subject of the CSP theory. info (temperatures, humidity, voltages, etc.) regarding The application of the process-oriented approach ANPR cameras and PC. It also gets the status of all into software development helps to map easily the CSP software components. The Status Manager Process sends processes to the OS processes or threads. Mentioned mentioned data via AM to the Web Server or logs it CSP approach decreases the execution time of the locally. implemented software, and as well improves the The WEB Server Process sends the live stream development of the flexible (open) architecture solutions. images and the results to the Web UI browser (operator). After the analysis of the requirements mentioned It is also responsible for online adjustment of the ANPR above, and taking into account the results achieved in system’s parameters. These parameters are exchanged be- [10] the ANPR system can be presented as one general tween Web UI browser and AM via WEB Server process. CSP process composed by the following (Fig 3.) pro- As can be seen the AM process is very compli- cesses: cated. That’s way it was additionally decomposed to set

79 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012 of sub-processes – each of them related to the specific activity or communication. - The Cameras, I/O Ports and DB Server can be interpreted as external hardware CSP processes which are connected to the ANPR System process via mentioned channels.

IDENTIFYING CSP CHANNELS AND MESSAGES

The arrows on the figure 3 represent the point- to-point CSP channels. Over these channels the Fig. 4. ANPR Manager Process sub processes. processes are exchanging messages. Channels pointing from AM to the surrounding processes are named ToXXX (where XXX corresponds to the targeted process) and channels named FromXXX are targeted to AM, for example FromOCR or ToOCR. Similar is the notation between other software and hardware processes. In general, the messages are of two types – messages with binary image data (in different formats and compressions) and text messages (in plain CSV or XML format). The length of the messages is variable Fig. 5. CSP composition of ANPR System. and depends on the size of the images or number of the message parameters. • UI Adapter Process communicates to the Web Server. It formats the live and result ER images and DECOMPOSITION OF ANPR MANAGER PROCESS sends them to the Web Server. It also exchanges and As mentioned above the AM process implements validates ANPR parameters data provided by the Web the main ANPR logic and it was functionally decomposed UI operator. (Fig. 4) to the next processes: • Software Updater Process is responsible for • ANPR Dispatcher Process is the main process online software (OS, ANPR) updates and for updates of the AM, which activates all other AM processes of the ANPR configuration files. depending on the raising events (messages) inside the • Input Output COM Ports Connector Process ANPR system. The ANPR Dispatcher and the external manages COM ports. for AM process - IH and OCR processes, have highest All mentioned above processes: - Rule Engine, priority. DB Connector, I/O COM Ports Connector, SW Updater • Evidential Record Builder Process creates and UI Adapter have the next (higher) priority in the encrypted evidential image of the captured vehicle system. The WEB Server and Status Manager processes containing plate number data, time, place and other have low priority. specific data. • Rule Engine Process sends ER data to remote DB Server or to COM ports in case some customer MAPPING THE CSP PROCESSES TO OS defined rules are matched. For example it sends ER to PROCESSES AND THREADS the DB Server in case the plate number matches one in After the decomposition the ANPR system was the White list and when the confidence level of the OCR represented of five CSP processes. The AM process it- results is greater than 80%. self was split to seven sub-processes controlled by one • DB Connector Process stores ER to the re- of them taking dispatching role. These processes are mote DB Server. It sends the ER as zip files using FTP mapped to OS-processes and threads according to their protocol.

80 A. Atanassov

importance. So OCR, Image Handler, ANPR Dispatcher ority. Last architecture was combination of above given processes and Evidential Record Builder must be architectures based of interacting processes with differ- executed with highest priority. Other processes forming ent priorities built of threads dispatched by one with AM have middle priority. The Web Server and the Status highest priority. Manager are the processes with low priority. Previous author’s investigations related to similar Next CSP record (Fig. 5) represents the prioritised software architectures [14] intended for parcels sorting parallel composition of all ANPR system processes. systems, confirms that first three architectures are not The sign || describes a parallel execution of two applicable for real-time application as ANPR system. <− processes or group of processes, and the sign || a parallel The Broker architecture is frequently used in embedded execution of two processes with higher priority than that systems and it reduces the context switches and inter- of the first (left-one) process. process communications because the Broker distributes A number of OS architectures [11, 12] were ana- the messages between the processes and controls the lyzed in order to select the appropriate OS process ar- common shared process memory. Last three architectures chitecture for ANPR system. Some of them are related are very suitable for real-time systems because the to one single OS process implementing sequentially the threads need less time for communications and share whole system activities or a number of interacting OS common memory inside the process. The architecture processes with equal priorities where each process with the dispatching thread corresponds fully to the CSP implements the activity of one CSP process. Other decomposition of the AM process. So the decision was architectures were related to a number of interacting taken to implement AM as one single OS process, built OS processes with different priorities where ach process of seven threads. These threads are related to the CSP implements the activity of one CSP process or a number sub-processes of the AM. The ANPR Dispatcher thread of interacting OS processes with different priorities, has high priority. It analyses the type of incoming coordinated by one special process – Broker. Next ar- messages from external processes and internal threads chitectures [13] are related to one single process com- and accordingly activates other objects signalling their posed by a pool of threads with equal priorities, where threads with corresponding messages/events. each thread implements the activity of one CSP process Finally, the ANPR System was mapped to five or to one single process composed by a pool of threads OS processes grouped in two priorities: high for IH, with different priorities, where each thread implements OCR and AM processes and, low for WEB Server and the activity of one CSP process. The activities of all Status Manager Processes. Inside the AM process there threads are dispatched by one thread with highest pri- are, also, two levels of priorities: highest for AM

Fig. 6. Result view from ANPR System in Web UI.

81 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

Dispatcher and ER Builder processes and middle for Press, 2007. the rest processes. 2. O. Martinsky, Algorithmic and Mathematical Prin- The decision to implement IH and OCR as ciples of Automatic Number Plate Recognition Sys- separated OS processes derives from the fact that in tems, University Brno, 2007, 5-18. most ANPR systems they are third party software 3. Recognising a new way to keep traffic moving, Si- (executables with known API) that can not be modified emens, Traffic April 2005, 1-12. or integrated inside the existing processes. 4. R. Bar-Hen, A Real-time vehicle License Plate Rec- The Web Server was implemented as a ognition (LPR) System, VISL, Technion 2002, Pa- multithreaded process and described in details in separate per at http://visl.technion.ac.il/projects/2003w24/ paper [10]. It is running inside the Apache Server [15] 5. http://www.ipcamerachina.com/CCTV-Camera-Se- installed on ANPR system PC. ries/LPR-ANPR-Camera/index.html. The real ANPR system (Fig. 6) ordered by 6. http://www.vitronic.de/en/traffic-technology/alpr/. Siemens is intended to work in Windows and Open 7. http://www.leutron.com/ Leutron Vision. Embedded OS environment and all threads’ and 8. http://x.htsol.com/index.php?action=1&k=4&sub=30 processes’ communications are based on the binary or Hi-Tech Solutions. text messages (see the table above) exchanged over TCP/ 9. C.Hoare, Communicating Sequential Processes, IP sockets. Prentice Hall, 1985, updated 2004. 10. A. Atanassov, F. Tomova, Web-Based Subsystem CONCLUSIONS For Tuning Of An Automatic Plate Number Rec- ognition System, Sixth International Conference, Presented in the paper process-oriented approach Challenges In Higher Education And Research approach ensures a better way for refining the software In 21st Century, Sozopol, Bulgaria, 2008, 399- architecture of the ANPR system by mapping efficiently 402. the CSP processes to corresponding OS tasks and 11. M. Barr, Introducing to Rate-Monotonic Scheduling; threads. The developed on the base of this approach Embedded Systems Programming; ANPR system is flexible and can be easily extended or www.embedded.com, 2001. reduced for different goals - speed control, traffic control 12. D. Kalinsky, 2001 Context Switch; Embedded (Fig. 6), highway tolling, lots control, etc. by adding or Systems Programming; Paper at www.embedded. discarding some processes or threads, using simple com; 2001. configuration files. In the future the ANPR system will 13. D. Pellerin, The Microsoft Win32 Programming be extended with additional module (OS process) Model: A Primer for Embedded Software Developers, responsible for on-line self update of the OS and other MSDN Microsoft Corporation, 1998. ANPR processes which will decrease dramatically the 14. A. Atanassov, Process-oriented Approach to Control time for maintenance. Interface Communications of Automated Parcels’ Processing Systems, Journal Automatics & REFERENCES Informatics 4, Sofia, 2009, 46-53. 15. T. Husted, Struts in Action Building web applications 1. A. Gordon, License plate recognition technology: with the leading Java framework, Manning innovation in law enforcement use, Thomson Gale Publications Co., 2003.

82 Journal of the University of ChemicalA. Atanassov, Technology L. Antonov and Metallurgy, 47, 1, 2012, 83-90

COMPARATIVE ANALYSIS OF CASE BASED REASONING SOFTWARE FRAMEWORKS jCOLIBRI AND myCBR

A. Atanassov, L. Antonov

University of Chemical Technology and Metallurgy Received 24 November 2011 8 Kl. Ohridski, 1756 Sofia, Bulgaria Accepted 10 January 2012 E-mail: [email protected]

ABSTRACT

This paper presents the comparison, in depth, of the capabilities of two Case-Based Reasoning (CBR) oriented software frameworks myCBR and jCOLIBRI for the development of predictive diagnosis and maintenance systems. Those frameworks were selected after detailed preliminary investigations [1] of the most popular noncommercial software systems for CBR as: myCBR, jCOLIBRI, CBR*Tools, CAT-CBR, CASPIAN. The evaluation of myCBR and jCOLIBRI includes the capacity to support the: R4 CBR circle; clusterization of cases, variety of used similarity functions, etc. Specific abilities to provide GUI, database support, required knowledge to work with the systems were also considered. Keywords: Case-Based Reasoning, predictive diagnosis, software comparative analysis, myCBR, jCOLIBRI.

INTRODUCTION ReCall, CBR-Works, ReMind[3], myCBR, jCOLIBRI, CBR*Tools [4], CAT-CBR [5] and CASPIAN [6]. The methodology of Case-Based Reasoning (CBR) for solving problems or tasks, based on already known Description of reasoning stages in CBR past decisions of same problems, is predicated on the The stages of reasoning in CBR systems, based hypothesis that similar problems have similar solutions on cases, are known as classical R4 cycle. Cases are the [2]. It is a part of artificial intelligence, connected to main object in CBR systems. They can be represented automation of reasoning based on cases, current problem as free text, in conversational type when each case is formulation, searching of one or more solved in the represented as a list of question and answers, or in past similar cases and adaptation of old solutions to the structural type when the cases are represented as a data current problem. base (case base). In this article deep comparative analysis of non- All structural cases are described as the pair commercial software frameworks myCBR and jColibri problem-solution [3, 7]. The problem p = (a , v ) is i i i is made. The main purpose of this analysis is to organized as a structure of attributes and values, determine the rate of applications of the software described by the attribute vector a = (a ,a , ,a ) and i i1 i2 ir frameworks for development of CBR-software platforms the value vector v = (v ,v , ,v ) i i1 i2 ir for the tasks of classic and predictive diagnostics. The solution s is represented as vectors, defined i The frameworks myCBR and jColibri are chosen by the specific tasks. In multidimensional supervised on detailed previous research, published in [1] of most control tasks, the decision includes two vectors s = (s , i pi distributed and used in practice commercial and non- p ), where the first vector sp = (sp , sp , ,sp ) consists ri i i1 i2 iq commercial software tools for CBR, such as Kate, of controllers sets on first hierarchical level, and the

83 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012 second pr = (pr ,pr , ,pr )– values of the target pa- • Required knowledge to use the software; i i1 i2 im rameters, corresponding to the sets. • Software support, updates and new versions; For solving an actual problem, the following 4 • Presence of graphical User Interface (GUI); main tasks of CBR R4 cycle are iteratively performed • Internal description of the cases (problems and (Fig. 1) [2, 6]: solutions); Retrieve – process of extraction of one (nearest • Type of similarity functions (SF); neighbor) or a group of cases (k-nearest neighbors) • Support of all parts (4R) of the CBR circle; having closest definition to the current problem. The • Support of different case-base DB (data bases); global similarity between the problems of these cases • Clusterization and indexation of the cases; (the new p and the one in the case base p ) is presented • Visualization of the cases; new j by following expression: • Application of software in specific scientific n areas = sim(pnew ,p j ) ∑ w isimi (pnew i ,p ji ) , • Support of interfaces to diagnostic or real-time = n i 1 systems; = and ∑w i 1 i=1 MyCBR is one of the most popular CBR software Where w is the weight of i-th attribute ≤ ≤ and platforms. It is a Framework with certain capabilities i 0 wi 1 sim(p , p ) is the local similarity between i-th attributes and limitations. The most popular version of myCBR is newi ji of the cases. a plug-in of open source ontology editor Protégé [8], but For global similarity measure the following there are available some WEB-based versions and metrics are most used: weighted Euclidian distance, integration into other software. myCBR is developed Manhattan’s metric, Humming’s metric, Tversky’s from German Research Center for Artificial Intelligence metric, Tchebishev’s metric, and minimum or maximum [9]. The platform has open source code written on Java metrics etc. and is accessible to all users. It can be easily modified Reuse – reuse of the solutions of chosen in the by the users depending on the purpose. The purpose of first step one or k-nearest neighbors. myCBR is to minimize the efforts to create CBR - When only one nearest case is chosen, the applications. For its normal use, without modifying the solution of the new problem s will be the solution of source code, no programming skills are required, but new the chosen case s . expertise in a specific CBR-developed applications.. Its NN - When k-nearest neighbors are chosen, the programming code is well documented. Work with the solution of the new case is calculated on the base of platform and its features are described in the manual, adaptation of k-nearest neighbors’ solutions. Revise – decision propriety and utility verifica- tion, made on the Reuse stage. This verification is mostly done by an expert or it is made based on simulation researches if there is a mathematical model available. Retain – saving (retaining) the new solution in the case base for future use, if it is successful

COMPARATIVE ANALISIS OF myCBR AND jCOLIBRI FRAMEWORKS The analysis of mentioned software frameworks is aimed to compare their capabilities for implementation of real-time and diagnostic CBR systems and takes into account following features: • Type and level and availability of software docu- mentation; Fig. 1. Classical Case-Based Reasoning R4 circle.

84 A. Atanassov, L. Antonov

Fig. 2. Definition of cases (classes), their attributes (slots), weights and global SF.

Fig. 3. Editing the local similarity function via GUI.

available at [9]. At present time no further develop- mum or minimum. On attribute level the SF can be ment and new versions of myCBR are available modified through the GUI, as shown in the middle of The framework myCBR supports description of Fig. 3 and they can be symmetrical, asymmetrical, step- cases with various attributes: numeric, character and type or smooth step-type, linear or polynomial. string, logical, class type, etc. The templates of the cases In myCBR the case and their attributes can be are generated as classes or subclasses with a number of created manually or automatically. The automatic attributes, called slots (Fig. 2.). generation of attributes (slots) is done during the import The CBR Cases are objects of the class described procedure of the Comma Separated Value /CSV/ file. by its attributes. Each attribute can participate in the Then to each column name from the CSV file is assigned class with its value and a weight that determines the an attribute with the same name. To each row of the file significance of the attribute in relation to others. the new case (instance of the class) is created in the Attributes a weight of zero (0) is not considered when case-base DB. With regard to maintenance the CBR 4R searching the case-base DB. cycle phases myCBR supports only Retrieve and Retain. Usually, case decisions have attributes with zero During the Retrieve phase all precedents extracted. They weight. In myCBR are given the opportunity to edit the are presented sorted by degree of similarity based on similarity functions (SF) on class level class (global SF) the chosen global SF. The Query to the case-base DB and on an attribute level (local SF). At the class level could be done on the basis of all or part of the attributes, the SF are: weighted sum, Euclidean difference, maxi- describing the case.

85 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

Fig. 4. Software architecture of jCOLIBRI.

Fig. 5. Definition of the Predictive Diagnostic class and its attributes.

86 A. Atanassov, L. Antonov

Table 1. Case-base data set.

On Retain phase myCBR allows to save the Query These three steps are described in details in as a new case, also to use an old case as a basis for new guidance’s and on-line lectures given in [4]. To create a Query. MyCBR is entirely based on GUI, providing a CBR application using the upper level no programmer’s ready-windows templates and forms for defining classes, knowledge is needed, only expert knowledge in the attributes, SFs, queries to the case-base DB, visualization specific scientific area of the application. The direct of found results and more. use of lower level modules allows development of myCBR does not work with external DB. It stores complicate CBR application, but this requires excellent the cases in text file or in XML file. That’s way it can programmer expertise and clear CBR system architecture not support the case indexation and clusterization. The definition. case can not be graphically presented in the GUI, but it jCOLIBRI is continuously developed in the last is possible to present the distribution of values of a 10 years and there are versions 1, 2 and 2.3 available. selected attribute for all cases in the database, as shown The use of last two versions requires knowledge in Java at right in Fig. 2. and the integrated development environment Eclipse. No interfaces to external systems and DB are There is a new version expected – jCOLIBRI Studio, available in myCBR. It is valid regarding the interfaces that will improve CBR systems with automatic to real-time or diagnostic systems. generation based on templates. jCOLIBRI supports full CBR cycle. At Retrieve jCOLIBRI is a framework for development of stage the nearest N cases are retrieved and there are 5 various CBR applications. It is Java-based and uses retrieval strategies, 7 selection methods and over 30 kinds JavaBeans technology for case representation and of SF in the spheres of text formatting and ontology. At automatically generation of user interfaces. This Reuse stage several methods for adaptation are available framework is developed by the GAIA artificial (direct proportion) and also in ontology. At Revise stage intelligence group in Complutense University in Madrid. methods for revision of cases are realized, as well The framework is built in two hierarchical levels – upper methods for new indexes (IDs) generation and methods and lower. The lower level (Fig. 4.) consists of library for decision making (preference elicitation). At Retain of classes (SW modules) for full R4 CBR cycle, also for stage there are methods for query retaining as a new definition of cases, attributes and connectors for access case. The maintenance algorithms such RENN, BBNR, to outer data bases. The upper level is a “black box” - etc. are also supported. jCOLIBRI allows retrieval form graphical interface, which allows non-complicated user clustered and indexed case bases and submits program CBR application generation based on lower level’s interfaces (connectors) to access text and XML files, as modules. Any application is generated in the following well standard and DL (descriptive logic)data bases. These three steps: interfaces can be used for diagnostic systems’ data bases - Pre cycle (attribute schema and local MF access. All CBR cases can be represented graphically. definition, selection of data base type, connectors for There are lots of CBR applications, developed on data base access selection); jCOLIBRI base: additional shells (abstract levels) for - Basic cycle (suitable retrieval methods selec- distributed CBR systems, statistical CBR systems, tion, global MF and methods for R4 tasks realization); multiagent supervisor systems [10, 11], systems for text - Post cycle (keeping results). files classification, and a lot of CBR recommender

87 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

Fig. 6. Definition of local SF (it can be graphically modified by the mouse).

Fig. 7. The results after the query to the case base DB. systems intended to trip, car type or restaurant choice, same figure the distribution of the values of the same and other of discrete-event type. attribute in all cases of the case base DB is given. Fig. 7 shows the results of the query to the case- Example with results base DB. In myCBR all cases are sorted in ascending way There are CBR applications developed with on the base of their proximity to the queried case. In the myCBR and for solving same problem – case processing estimations of the proximity the local and global SF are given in Table 1. All cases concern diagnostics (work taken into account. Figs. 8 and 9 show attribute defini- behavior) of drill machine in mine industry. Columns tion in analogical CBR application, made with jCOLIBRI, A, B, C and D in table 1 are the problem attributes of and the results of the same query as shown at Fig. 6. the cases and columns E and F – the decision attributes. In jCOLIBRI the similarity function for attributes All data is processed using the methodology described having floating point values is only equal. In order to in [12]. First application is based on myCBR and second avoid inappropriate results during the Retrieval phase one on jCOLIBRI. the approach to present all floating point values as inte- Fig. 6 shows graphical modeling of the similar- ger ones is used. So the values of all integer attributes ity function of asymmetric polynomial type of the se- are multiplied by 100 and the most sophisticated simi- lected attribute (Hole Profile). On the right side of the larity functions defined for integers are used.

88 A. Atanassov, L. Antonov

Fig. 8. Definition of the attributes in jCOLIBRI ver. 1.1.

Fig. 9. The results after the query to the case base DB. Conclusion and future work jCOLIBRI can be used as a basis for complex CBR applications development with full CBR R4 cycle, myCBR platform can be used for non-complex using various data bases. Development of this kind of CBR applications development with partial CBR R4 applications however requires excellent programmer cycle and with small number of cases in text file. For knowledge, time for requirements definition, CBR application development no time for programming development of software architecture, complicated is needed; only for case configuration. MyCBR is not graphical user interface, data base configuration and time suitable for application with large number of attributes for implementation, test, adjustment and verification. with text solution, especially when they must be visu- Based on the examples, given above, it is obvi- ally presented in one window. ous, that myCBR interface overmatches jCOLIBRI‘s and

89 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

gives more options for weights and SF type modifica- REFERENCES tion of attributes and cases. This is of great importance for query adjustment and refining to the case base. 1. K. Boshnakov, L. Antonov, A. Atanassov, Comparative Suggestions for myCBR usage – the Java code analysis of software implemented method of prece- of myCBR to be expanded to work with hyperlinks and dents in the tasks of diagnosis, Anniversary Scientific data bases. In case it is not possible an external program Conference “40 Years Department of Industrial Au- converter to be developed. It can be intended to read tomation”, University of Chemical Technology and external data base and to convert all cases in the format Metallurgy”, Sofia, 2011, 89-93. (CSV files), used in myCBR, as well to ensure back-way 2. Y. Avramenko, A. Kraslawski, Similarity Concept for conversion. Case-Based Design in Process Engineering, Computers Suggestions for jCOLIBRI usage – to be and Chemical Engineering, 30, 2006, 548-557. developed an application, that uses one or more data 3. S. Pal, S. Shiu, Foundations of Soft Case-Based bases; one for cases and their solution, and other Reasoning, Wiley-Interscience, 2004. (operative DB) for on-line data, that could be available 4. Case-Based Reasoning Framework jCOLIBRI, http:/ for other programs for filtration, statistical or other data /gaia.fdi.ucm.es/projects/jcolibri/. processing of the diagnosis object or system. 5. Component Architecture Technology for developing Suggestion for development of own CBR software CBR systems, CAT-CBR, http://www.iiia.csic.es/ application –which can support groups of data bases - Projects/cbr/cat-cbr/ . one for the cases and the solutions, and other one - for 6. CASPIAN, www.aber.ac.uk/~dcswww/Research/ on-line data of the diagnosis object or system. Also the mbsg/cbrprojects/getting_caspian.shtml development of specific software intended for input/ 7. CBR*Tools, www-sop.inria.fr/a[xis/cbrtools/ output, for case retrieval from case base DB, for filtration, usermanual-eng/Introduction.html#OOFramework adaptation, etc. is recommended. The advantages of data 8. Protege, http://protege.stanford.edu/. bases are that they can keep complex cases in tables with 9. myCBR, http://www.mycbr-project.net/. relations to other tables with graphical and/or picture 10. K. Boshnakov, C. Boishina, M. Hadjiiski, Multiagent information or relations to tables with lectures, that fault-tolerant supervising control of wastewater contain decisions and recommendations for solving treatment plants for wastewater, International Con- specific problems. The last mentioned is not realized in ference, Automatics and informatics’11, Bulgaria, considered in this article frameworks, neither in the CBR Sofia, 2011. software analyzed in [1]. 11. M. Hadjiski, V. Boishina, Enhancing Functionality of Complex Plant Hybrid Control System Using Case- Acknowledgements Based Reasoning, Intelligent Systems (IS), 2010 5th The studies were funded by the Fund “Scientific IEEE International Conference, 7-9 July 2010, Research” (F”NI”) under the Ministry of Education, London, 25-30. Youth and Science under the project “Predictive 12. D. Tenchev, G. Kondev, Total Maintenance of Equip- maintenance of technological equipment based on diag- ment, MJ Publishing Technical University of Sofia, nosis and risk analysis” ¹ DVU-10-0267/10. 2006.

90 Journal of the University ofSt. Chemical Georgieva, Technology T. Nedeltcheva and Metallurgy, 47, 1, 2012, 91-96

DETERMINATION OF THE OXYGEN STOICHIOMETRY OF Y1Ba2Cu3OY SUPERCONDUCTORS WITH DIFFERENT OXYGEN CONTENT

St. Georgieva, T. Nedeltcheva

University of Chemical Technology and Metallurgy Received 04 November 2011 8 Kliment Ohridski, 1756 Sofiaÿ, Bulgaria Accepted 10 January 2012 E-mail: [email protected]

ABSTRACT

The spectrophotometric method using the measurement of the absorbance of an I --starch compound is applied 3 for the determination of the oxygen stoichiometry of high temperature Y Ba Cu O (y=6.5+ä) cuprates with oxygen 1 2 3 y content from 6.6 to 7.0. In order to measure the absorbance values of the solutions precisely some changes in the procedure are made. The oxygen stoichiometry of the samples is determined by the X-ray method as well. Keywords: spectrophotometry, YBCO superconductors, oxygen determination, X-ray method

INTRODUCTION two absorbance values would be higher than two (for example if ä is 0.1, the ratio is 10) and the precision of The determination of the oxygen content of high the absorbance measurements would be poor [24]. temperature Y Ba Cu O (y=6.5+ä) superconducting Raman and X-ray methods are also used for the 1 2 3 y ceramics is important because the superconducting determination of the oxygen stoichiometry of the high properties are correlated with the oxygen coefficient in temperature superconducting ceramics [22, 23] The XRD the range of “y” from 6.60 to 6.99 [1]. Various methods method uses a mathematical relation between c-axis have been developed for the determination of the oxygen parameter of the orthorhombic cell and the oxygen content stoichiometry of Y B Cu O (YBCO) samples [2 – 23]. “y” of the superconducting ceramic. Such a mathematical 1 2 3 y The most attractive among the “wet” methods is the relation “oxygen content – physical parameter ” has been spectrophotometric one, which is based on measuring derived from Inorganic Crystal Structure database (ICSD) the absorbance of an I --starch compound [20, 21]. It data for Y Ba Cu O ceramics [23]. 3 1 2 3 y does not require both measuring the sample mass and The aim of this work is to apply the iodine-starch calibration of the analytical function. The non- spectrophotometric method [20] for the determination stochiometric oxygen coefficient “ä” is calculated as a of the non-stoichiometric coefficient “ä” of Y Ba Cu O 1 2 3 y ratio of the absorbance values of two solutions prepared ceramics with oxygen content from 6.6 to 7.0. from the solution of a same sample and measured against a comparative solution that contains also a volume of EXPERIMENTAL the sample solution. The method proposed is applied to Sample preparation samples with high oxygen content. The values of the Synthesized sample non-stoichiometric coefficient ä of the samples are close A solid-state reaction was used to prepare the to 0.5 and the ratio between the two absorbance values Y Ba Cu O sample [25, 26]. The triple heat treatment 1 2 3 y is about two. If however the non-stoichiometric oxygen regime with intermediate grindings was used. The first coefficient values are lower than 0.5, the ratio of the step includes mixing and milling of appropriate amounts

91 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012 of Y O , CuO, BaCO , and calcination in flowing oxy- to the sample and blank solutions. After this proce- 2 3 3 gen at 900 ºC for 21 h. The second step of the heat dure, 6 ml of the sodium hydroxide solution were added treatment was conducted at 930oC for 21 h in the same to each of the beakers and the stirring of the solutions atmosphere, followed by annealing at 450oC for 2 h. was stopped. Two volumes of 1.00 and of 5.00 ml were The last step (third) started with grinding and pressing taken from the sample solution and transferred into three of the powder in pellets followed by sintering at 940oC measuring flasks of 10 ml (see Table 1). Two volumes for 23 h, slow cooling to 450oC and holding at that of 4.00 ml of the blank sample were added to the flasks temperature for 23 h. 2 and 3, which contain 1 ml of the sample solution. Thereafter, 2 ml of glycine solution were added to flasks Annealing samples 1 and 3 (flask 1 contained 5.00 ml of the sample To obtain samples with lower oxygen content, solution) and 2 ml of distilled water were added to flask pieces of the synthesized superconducting sample were 2. The solutions were homogenized and remained in a grinded and annealed at 460±5oC in flowing argon for dark place for 10 min. After that, 1 ml of the 0.8 % various times, then quenched to room temperature [23]. starch solution was added to each flask drop by drop while shaking, the flask and the solutions were diluted Sample analysis by the spectrophotometric method up to the mark of 10 ml with distilled water. After 5 Solutions and apparatus min the absorbances were measured at ë=545 nm and The following reagents were used: 0.1 mol l-1 the absorbances of the solutions in flasks 1 and 2 were hydrochloric acid; 0.1 mol l-1 sodium hydroxide solu- read against the solution in flask 3. It is recommended tion; sodium acetate-acetic acid buffer (C = 0.08 to prepare two parallel sets of solutions and the NaAc mol l-1 ; pH = 4.9 ±0.1); 2 mol l-1 glycine solution; 0.8 coefficient of the active oxygen to be calculated by the % solution of starch in water (v) : glycerine (v) mixture mean absorbances. (1:1) – the solution is stable for several weeks; potassium Sample analysis by the X-ray method iodide (p.a. Merck); nitrogen gas (99.9 %). Powder X-ray diffraction (XRD) data were Two sets of beakers and funnels were used for recorded at room temperature using Bragg-Brentano sample dissolution, the one for the YBCO sample and geometry with Co Ká radiation. The diffractograms were the other for the blank sample [27]. indexed using DICVOL [28] and the obtained unit cell The absorbance was measured by a single beam parameters were refined using the X’pert Highscore Spekol 11 spectrophotometer (Carl Zeis, Jena) using program. The peak shape was described by a pseudo- cells with path length of 1 cm. Voigt function. The background level was defined by a Procedure polynomial function. For each diffraction pattern, the A mass of about 1.5 - 2.0 mg of a bulk sample, scale factor, the peak asymmetry and the unit-cell powdered and homogenized in advance, was placed in a dimensions were refined while the instrumental zero hemispherical glass container and put on the bottom of shift was kept constant. the one of the beakers. Two measures of 2.5 g KI were placed in the beakers. A volume of 6 ml of the RESULTS AND DISCUSSION hydrochloric acid was introduced in each of the funnels. The oxygen from the beakers and the solutions was The spectrophotometric procedure is close to that removed by purging with nitrogen gas for 10 min. The given in [20]. Difference exists only in the sample volumes hydrochloric acid solutions were introduced to the taken to prepare the solutions for the absorbance beakers and the potassium iodide was carefully measurements. In the procedure proposed two volumes dissolved before the sample made contact with the of 1.00 and one of 5.00 ml were taken (Table 1) inståad solution. Then the nitrogen gas was stopped, the sample of two volumes of 2.00 ml and one of 5.00 ml according was poured from the glass container and dissolved by to the procedure given in [20]. The non-stoichiometric stirring. The funnels were removed and volumes of 9 coefficient ä was calculated by the formulas: A ml of the sodium acetate-acetic acid buffer were added ä= A1 [20] and ä=0.375 1

A2 A2 92 St. Georgieva, T. Nedeltcheva

Table 1. Preparation of the solutions for the absorbance measurements (procedure 1:1:5)

Volumes, ml Number of the Blank flask Sample sample* Glycine d. H2O Starch

Flask 1 5 - 2 - 1

Flask 2 1 4 - 2 1

Flask 3 comparative 1 4 2 - 1 solution

*The blank sample passes through all stages of the procedure, simultaneously with the sample for analysis

All prepared Y Ba Cu O samples are monophases known and the proposed procedures cannot be distin- 1 2 3 y with 123 orthorombic structures (Fig. 1). The oxygen guished. With regard to the ratio of the absorbance val- content of YBCO samples determined by the proposed ues there is a difference. In the procedure 2:2:5 it varies procedure 1:1:5 is from 6.68 to 6.92 (Table 2). The from 2.4 (ä=0.42) to 5.9 (ä=0.18), while in the proce- decrease of the non-stoichiometric oxygen coefficient ä dure 1:1:5 the absorbance value ratio is from 1.1 one to as a function of the annealing time is given on Fig. 2. 2.1. Thus, in the proposed procedure the absorbance As seen, the process is very rapid during the first 12 h. values of both solutions are close and they would al- After the 12-th hour the oxygen decrease is time ways be measured with good precision. consuming. The values of the unit cell c-parameter of all The samples with the lowest (ä=0.18) and with samples we determined. They are given in Table 2. Us- highest (ä=0.42) oxygen content we analyzed by the ing the known equation y = 75.250-5.856c [23] the val- procedure 2:2:5 [20] as well. The results are given in ues of y we calculated and listed in Table 2 also. The Table 3. It is seen that the ä-values obtained by the difference Äy between y-results obtained by the spec-

0,45

0,40

0,35

0,30

0,25 Non-stoichiometric coefficient 0,20

0,15 0 20406080100 Annealing time, h

Fig. 1. XRD patterns of starting (0 h) and annealed in flowing Fig. 2. Decrease of the non-stoichiometric oxygen coefficient argon (6 – 96 h) YBa2Cu3Oy samples. ä of YBCO samples as a function of the annealing time.

93 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

Table 2. Values of the c-axis of the cell and the oxygen coefficient, y.

Spectrophotometric XRD - method

Sample method

oxygen oxygen c-unit cell oxygen Annealing coefficient coefficient parameter time, [h] coeficient ó * y y 0 0.418 6.918 11.676(5) 6.875 0.043

YBCO 6 0.272 6.772 11.696(5) 6.758 0.014 12 0.204 6.704 11.699(6) 6.741 -0.037 96 0.178 6.678 11.713(6) 6.659 0.019

*Each result is a mean value of at least three parallel determinations. The confidence interval with P=0.95 for each sample is ± 0.004.

Table 3. Values of the absorbance and the ä-coefficient of two YBCO samples.

YBCO m (mg) Procedure 2:2:5 Procedure 1:1:5 sample With A1 A2 Absorbance A1 A2 Absorbance ratio ratio

1.35 0.272 0.663 2.4 0.410 1.64 0.337 0.796 2.4 0.423 1.52 0.326 0.780 2.4 0.418

1.28 0.364 0.318 1.1 0.429 highest 0.372 0.326 1.1 0.428 oxygen content 1.32 0.351 0.313 1.1 0.421 0.361 0.323 1.1 0.419 1.67 0.455 0.418 1.1 0.409 0.471 0.434 1.1 0.407 1.47 0.426 0.386 1.1 0.414 1.1 0.414 0.371 0.415 1.86 0.543 0.491 1.1 0.415 1.96 0.152 0.879 5.8 0.173 2.10 0.168 0.928 5.5 0.181 lowest 2.01 0.164 0.970 5.9 0.169 oxygen content 2.15 0.175 0.946 5.4 0.185 2.13 0.230 0.483 2.1 0.178 2.08 0.228 0.475 2.1 0.180 2.30 0.272 0.582 2.1 0.175 2.01 0.213 0.441 2.1 0.181

94 St. Georgieva, T. Nedeltcheva

trophotometric and the XRD-methods are also presented mining the oxygen content of samples doped with (Table 2). Having in mind that a deviation of 0.005 in Fe and Co, Physica C, 270, 1996, 349. the value of the c-parameter causes deviation of 0.034 9. W.M. Chen, J. Chen, X. Jin, A modified function of in the calculated y-value one can conclude that the dif- the formula of double iodometric titration, Physica ference between y-values is not significant and system- C, 276, 1997, 132. atic errors are absent. 10. W.M. Chen, X.S. Wu, J.F. Geng, J. Chen, D.B. Chen, X. Jin, S.S. Jiang, An accurate method of iodometric CONCLUSIONS titration to measure copper valence of high-T c superconductors, J.Supercond, 10, 1997, 41. With the proposed changes, concerning the 11. K. Kishio, J. Shimoyama, T. Hasegawa, K. Kitazawa, volumes of the samples taken to prepare two solutions K. Fueki, Determination of Oxygen Nonstoichio- for absorbance measurements, the non-stoichiometric metry in a High-T Superconductor YBa Cu O , Jpn. c 2 3 7-ä oxygen coefficient of Y Ba Cu O samples with different J. Appl. Phys., 26, 1987, L1228-L1230. 1 2 3 y oxygen content can be successfully determined. 12. K. Fueki, Y. Idemoto, H. Idhixuks, Oxygen nonstoichiometry of a Y(Ba Sr ) Cu O 1"x x 2 3 7"ä REFERENCES superconductor, Physica C, 166, 1990, 261. 13. L. Liu, C. Dong, J. Zhang, H. Cheng, L. Chen & all, 1. S. Kambe, Y. Murakoshi, R. Sekine, M. Kawai, K. A simple volumetric method for oxygen content Yanada, S. Oshima, K. Okuyama, Dramatic changes determination in high-T doped YBCO compositions, c in the Bi and Cu valences of Bi 2201 phase annealing Physica C, 383, 2002, 17. in absence of oxygen, Physica C, 178, 1991, 71. 14. P. Richter, M. Kreb, E. Mohler, G.H. Roskos, G. 2. D.C. Harris, T.A. Hewson, Determination of Cu3+/ Jakob, H. Adrian, Optical determination of the Cu2+ ratio in the superconductor YBa Cu O , J. oxygen content of YBa Cu O thin films by IR 2 3 8-x 2 3 6+x Solid State Chem, 69, 1987, 182. reflectance and transmittance measurements, Physica 3. A.I. Nazzal, V.Y. Lee, E.M. Engler, R.D. Jacowitz, C, 366, 2001, 63. Y. Tokura and J.B. Torrance, New procedure for 15. M. Karppinen, H. Yamauchi, S. Tanaka, An Improved determination of [Cu-O]+p charge and oxygen content Coulometric Titration Method for Oxygen Content in high T copper oxides, Physica C, 153-155, 1988, and Valence determinations in Bismuth and Lead c 1367-1368. Cuprates, J. Solid State Chem., 104, 1993, 276. 4. M. Karppinen, L. Niinisto, Determination of the 16. R. Chavdarova, T. Nedeltcheva, L. Vladimirova, oxygen content in the superconducting YBa Cu Determination of non-stoichiometric oxygen in 2 7-ä phase, Supercond. Sci. Technol., 4, 1991, 334. YBCO superconductors by coulometric titration, 5. T. Nedelcheva, A. Stoyanova-Ivanova, L. Kostadinova, Anal. Chim. Acta, 353, 1997, 325. I. Ivanova, Determination of Cu(III) and 17. C.J. Barbour, B.L. Doyle, S.M. Myers, Measurement Cu(II)+Cu(III) in superconducting copper ceramics, of the oxygen content in high-T superconductors: c Fresenius J.Anal. Chem., 342, 1992, 196-198. Enhanced resonant ion-scattering analysis, Phys. Rev. 6. J.R. Min, J.K. Liang, H. Jin,C. Wang, H. Chen, Y.D. B, 38, 1988, 7005. Yu, C. Dong, G.H. Rao, A new family of Pb-based 18. T. Nedeltcheva, Anal. Chim. Acta, Determination 1222 cuprates Pb(Sr,La) Ln Cu O (Ln=Gd, Dy, Eu, of oxygen stoichiometry in YBCO superconductors 2 2 2 z and Pr), Physica C, 249, 1995, 196. by spectrophotometry, 312, 1995, 223. 7. A. Fukuoka,M. Karppinen, N. Seiji, J. Valo, A. 19. T. Nedelcheva, L. Vladimirova, Anal. Chim. Acta., Kareiva, L. Niimisto, M. Leskela, N. Koshizuka, H. Spectrophotometric determination of oxygen Yamauchi, Wet chemical determination of the oxygen stoichiometry in YBCO superconducting bulk content in YBa Cu O samples synthesized by various samples, 437, 2001, 259. 2 4 z methods, Supercond. Sci. Technol., 8, 1995, 673. 20. T. Nedeltcheva , St. Georgieva, L.Vladimirova and 8. W. Chen, W. Hong, J. E .Geng, X.S. Wu, W. Ji, L.Y. A. Stoyanova-Ivanova, Talanta, Increasing the sen- Li, L. Qui, X. Jin, Iodometric titration for deter- sitivity of the spectrophotometric determinations of

95 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

the oxygen content in YBCO superconducting 1969, (in Russian). samples using the I "-starch compound, 77, 2009, 25. S. Terzieva, A. Stoyanova-Ivanova, K. Zalamova, V. 3 1745–1747. Mikli, Ch. Angelov, V. Kovachev: Morphology of 21. A. Stoyanova-Ivanova, St. Georgieva, T. Nedeltcheva, Y Ba Cu O and Y Ca Ba Cu O bulk samples de- 1 2 3 z 0.7 0.3 2 3 z L. Vladimirova and S. Terzieva, Spectrophotometric pending on Ca-subtitution, J. Optoelectron. Adv. M., determination of oxygen stoichiometry in 7, 1, 2005, pp. 477–480. modification RE(rare earth elements)Ba Cu O 26. C.N.R. Rao, R. Nagarajan and R. Vijayaraghavan, 2 3 z superconducting samples, JOAM, 1, 2009, 476-478. Synthesis of cuprate superconductors, Supercond. Sci. 22. S. Degoy, J. Jimenez, P. Martin, O. Martinez, A. Tech., 6, 1993, 1–20. Prieto, D. Chambonnet, C. Audry, C. Belouet and J. 27. A. Stoyanova-Ivanova, T. Nedeltcheva, L. Vladimirova, Perriere, Oxygen content of YBaCuO thin films, Spectrophotometric determination of oxygen con- Physica C, 256 1996, 291–297. tent in calcium substituted RBCO (R=Eu, Gd, Er) 23. P. Benzi, E. Bottizzo and N. Rizzi, Oxygen determi- superconductors, Cent. Eur. J. Chem., 3, 2005, 432- nation from cell dimensions in YBCO superconduc- 440. tors, Journal of Crystal Growth, 269, 2004, 625– 28. A. Boultif and D. Louer, Indexing of powder dif- 629. fraction patterns for low-symmetry lattices by the 24. V.F. Barkovskii, V.I. Ganopolskii, Differential spec- successive dichotomy method, J. Appl. Cryst., 24, trophotometric analysis, Chimia Publisher, Moscow, 1991, 987–993.

96 Journal of the UniversityV. Bardarov,of Chemical M. TechnologyShishenkov, andK. Bardarov Metallurgy, 47, 1, 2012, 97-102

APPLICATION OF PREPARATIVE SIZE-EXCLUSION CHROMATOGRAPHY (SEC) FOR SEPARATION/CLEAN-UP OF PLATELETS FROM BLOOD PLASMA BEFORE THEIR CHEMICAL INVESTIGATION

V. Bardarov1, M. Shishenkov1, K. Bardarov2

1Military Medical Academy Received 28 November 2011 3 Georgi Sofiisky blvd. Sofia Accepted 12 January 2012 2 St. Kl. Ohridsky University, 1 James Bouchier Ave. 1164, Sofia

ABSTRACT

Platelets have important functions in haemostasis and their investigation is needed. Some of their characteristics as platelets concentration in blood, their volume, etc., are conventional parameters, measured by automated laboratory analyzers. When their chemical investigation is needed, a procedure for their separation and clean-up is required, because their “porous” structure and active surface retains components from their environment being obstructive in obtaining data about their chemical composition. A procedure for separation and clean-up of platelets by preparative size-exclusion chromatography is described in this work. 350x16 mm chromatographic column with Sepharose 4B as stationary phase and CDP/Heparine/Dipyridamol solution as eluent were used for platelets separation from blood plasma, being previously set free of erythrocites and leucocites by centrifugation. Fraction 17-19 ml was collected, platelets were sedimented by centrifugation at 6000 rpm for 5 min. The upper liquid was discarded and the platelets were suspended in a convenient media for their chemical investigation. When in the process of chromatographic separation a light-scattering detector is used, some additional information for platelets can be collected as their size-distribution and hydrodynamic properties together with precise collection of fraction containing pure platelets. Using the CDP/Heparine/Dipiridamol solution as eluent supports platelets in vital form and preventions changes in their chemical composition and physical state. The procedure described permits preparation of platelets in pure, native, not-destructed and convenient for their chemical investigation form. Keywords: platelets, preparative liquid chromatography, SEC, sample preparation.

INTRODUCTION are generated from the cytoplasm of a mature megakariocites of the red bone marrow and being de- Platelets are blood formed elements with 2-3 µm tached from it their membranes are formed. With these diameter and discoid shape. They do not have nuclei, membranes platelets interact with their environment but have peculiar formations called alpha granules and (blood plasma) and in between. Platelet’s membrane dense bodies [1] and their membranes posses interesting has specific biochemical and physiological properties physiological properties besides surrounding the cell and features [6, 7]. That is why the biological function- cytoplasm and the organelles. Their blood concentration ality of these cells is connected with some phenomena is 130-360x109/l and their lifetime in blood circulation of their plasma membranes [8-10], most likely also with is seven to ten days [2]. the chemical structure and composition of their con- Platelets are considered to lack DNA [3], as well structive elements – phospholipids. as RNA, endoplasmic reticulum or ribosomes [3] Before investigation of the composition of plate- despite some contrary author opinions [4, 5]. These cells lets’ membranes one has to acquire intact and clean

97 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012 insolated cells. Platelets should be possibly free from parative separation of platelets from the blood plasma the other blood elements but also from any plasma components. The column is made by suspending of the components that might have got tightly hold on them. gel in degassed HPLC-grade water, homogenization of The highly folded platelet membrane resembling a the gel by ultrasonification and introduction of the “sponge” as well as any electrostatic interactions with suspension in the column in portions while water flows the media components hamper the platelets purification out under gravitation. The column is further made tight and the popular gravitational methods are not always by letting 100 ml HPLC-grade water to flow through it. enough effective for the goal of purification [11] and An “Olympus CX-41” phase contrast microscope platelets preservation. Besides clean surface, the fit for is used for observation of the cells in blood plasma and purpose of further examination of platelets isolation platelets pure preparations. procedure should keep them in their native shape and A SIGMA 1-6 laboratory centrifuge is used for should not disrupt them or provide conditions for the gravitational isolation of platelets. aggregation. For this reason the make-up of the media that is used in the clean-up is a particularly important Chemicals factor. The relatively small size of these cells and their „CPDA-1" (“JMC Singapore PTE” Ltd), consist- narrow size distribution make it possible to purify them ing of 0,299 g citric acid (anhydrous), 2.63 g sodium with chromatographic means. There are some techniques citrate (dihydrate), 0.222 g monobasic sodium phos- that have shown positive results for purification of blood phate (monohydrate), 3.19 g dextrose (monohydrate) and cells in the literature [12]. 0.0275 g adenine in 60 ml sterile water; Optimising the conditions for isolation and Platelets-stabilization solution, consisting of 10 % purification of platelets from blood plasma with their CPDA-1, 50000 U/L lithium heparine (“Calbiochem”, entirety preservation, which is the subject of this work, Merck, Germany), 80 µg ml-1 dipyridamole (“Calbiochem”, is the first step on the way for investigation of their Merck, Germany), 0.9 % NaCl in HPLC grade water. chemical composition. Biological material EXPERIMENTAL Human venous blood samples from anonymous Equipment patients, taken in vacutainers with EDTA as an antico- A hematology analyzer (Sysmex XE Series), mea- agulant, was provided from the routine work-stream of suring scattered light and a hematology analyzer (Diagon a clinical laboratory. Samples are chosen on random D-cell 60), measuring conductivity, are used for plate- principle but in correspondence with the experiment lets number count in blood plasma and in the fractions requirements. Platelets are separated in the form of collected from the preparative gel-chromatography. “platelet rich plasma” (PRP) not more than 60 minutes Olympus AU 640 automated clinical chemical after blood collection and the different plasma samples analyzer is used for total protein quantitation in fractions were mixed together. Samples were not kept in fridge. collected by preparative gel-chromatography, according Procedure a validated photometric method. Erythrocytes and leukocytes are gravitationally A Zetasizer Nano DLS unit (“Malvern separated from the blood plasma by centrifugation of Instruments”), equipped with 70 µl quartz flow cell is the blood at 190xG for 20 minutes. Upper liquid layer, used as a detector for on-line measurement of eluting after centrifugation, PRP consists of platelets and plasma particles and some other big biomolecules from the molecule species in their natural solution. chromatographic column, and for measurement of their Platelets isolation and clean-up was performed size distriburion. with preparative gel-permeation chromatography. A 350x16mm glass chromatographic column Before sample introduction in the (“Pharmacia Fine Chemicals”, Uppsala, Sweden) with chromatographic column, the last was saturated with floating frit, filled with a poly-succrose gel “Sepharose the platelets-stabilization solution used also as mobile 4B” (“Pharmacia Fine Chemicals”) is used for the pre- phase in the chromatographic process.

98 V. Bardarov, M. Shishenkov, K. Bardarov

After saturation 2 ml aliquot of PRP was intro- tive GPC with “on-line” detection, and monitoring of duced into the floating cup of the column. The elution particles and big molecules in the eluate, is presented of the sample was performed with the stabilization so- on Fig. 1. Curve 2, showing the light scattering of the lution at a flow rate of 0.7 ml min-1 and 1 ml fractions eluate from the column, registers two peaks at RT 17- were collected in marked tubes. Immediately after col- 20 min and at 40-47 min. lection the number of platelets, size distribution and The particle size evaluation (curve 1) shows a the total protein concentration were measured in each small number of large size particles in a narrow band at fraction. RT 15-16 min, and a wide band of lower size particles, When on-line monitoring with use of Zetasizer eluted after 17th min. Curve 1 shows small size particles, Nano DLS as a detector, the corresponding probably protein molecules, presented in the band 40- chromatograms were recorded and the relevant 47 min, registered by light scattering intensity elevation parameter characteristics of eluted particles were in curve 2 as well. registered as well, presented in Figs. 1 and 2. Fig. 2 represents the volumes of size distributed particles in the whole sample. There are five groups of RESULTS AND DISCUSSION different by size particles: Particles with size 2100-2700 nm in diameter corresponds to the size of platelets. Platelets number and total protein content Particles with largest relative volume and smallest size measurement in PRP and in preparative-GPC fractions (100-200 nm in diameter), are probably proteins, cor- An elution curve for platelets and proteins of responding to the 40-47 min band in curve 2 on Fig. 1. platelets rich blood plasma (PRP), obtained by prepara- The largest size particles with around 3000 nm in di-

Fig. 1. Gel-permiation chromatogram of a 2 ml sample of PRP with “on-line detection by Z-Sizer with 70 µl flow cell as detector.

Fig. 2. Size distribution of particles and large molecules in total eluate of ORP with volume of size-homogenius groups.

99 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

Fig. 3. Microscope picture of particles/cells in PRP fractions (a – RT=16 min; b – RT=18 min). ameter, are probably single small erythrocites, corre- data obtained with higher sensitivity, is shown on Fig. 5. sponding to peak at RT 16 min on curve 1 on Fig. 1. Platelets, eluted in fraction 16-20 ml, in a relative narrow The other two groups of particles with size around 1100 chromatographic peak, are well separated from proteins, and 1800 nm, are unknown particles, presented in the eluted in fraction 37-55 ml, and the size distribution of sample. platelets in the fraction 17-19 ml shows more narrow For obtaining of additional information about band and average platelets volume 7.0 fl, compared to the particles in band 15-20 min (Fig. 1), two fractions size distribution of platelets in PRP with average (15-16 min and 16-18) were collected and examined platelets volume 9.6 fl (Fig. 6). The recovery of platelets, under microscope. Figs. 3a and 3b represent their obtained by preparative GPC is 22-45 % (measured 6 microscope images, showing some erythrocites in the times), which is relatively low, but the platelets obtained first fraction (corresponding to the large size particles are more homogeneous, free of outside proteins and in curve 1 on Fig. 1), and presence of bacteria with non aggregated. The column is free of plasma compo- different sizes (corresponding to the wide band 16-35 min in curve 1 on Fig. 1), covering and tailing the band of platelets. The presence of bacteria in the sample is due to the long period of stay of the PRP (approx. 36 hours) before chromatographic separation. A microscope picture of a chromatographic fraction (16-20 min) of a freshly prepared and processed PRP (Fig. 4) shows free of bacteria platelets fraction. It was not possible to obtain a second “on-line” detection of this sample, due to Zeta-sizer usage restrictions. Due to the low sensitivity of Zeta-sizer, used as a light-scattering detector (curve 2 on Fig. 1), fractions of 1 ml from the SEC column were collected. The concentration of platelets were measured by an automated Fig. 4. Microscope picture of fraction (RT=18-19 min) of hematology analyzer, and the total protein content - by a freshly collected blood, followed by immediate PRP biochemical analyzer. The chromatogram, built from the preparation and fractionation.

100 V. Bardarov, M. Shishenkov, K. Bardarov

Fig. 5. Chromatogram of preparative chromatographic separation of platelets by “of-line” detection. Platelets number and total protein content are measured in collected fractions.

Fig. 6. Average distribution of platelets by volume, in freshly prepared PRP (a) and in fraction 16-19 ml after preparative GPC (b). nents after washing with 60 ml eluent, and it can be lets and could be the first step of further investigations used for next platelets separations. on platelets structure and composition. The procedure The platelets from the fraction after preparative is enabling to be automatic, with reduction of time and GPC separation can be deposited by centrifugation at efforts. 6000 rpm for 15 min, and after discard the elution solution can be suspended in convenient media for their REFERENCES investigation 1. R. Baklaja, M.C. Pešic, J. Czarnecki, Hemostasis and CONCLUSIONS hemorrhagic disorders, Fermentation-Biotec GmbH, Bad Harzburg, 2008. The procedure described for preparative GPC 2. M.B. Zucker, The value of the blood platelet in platelets separation/purification should be carried-out hemostasis and as a scientific tool, Trans. N.Y. Acad. at sterile conditions. It is obvious that the treatment Sci. Series 11. 36, 1974, 561-568. described allows, even with low recovery, to obtain 3. A.J. Marcus, Journal of Lipid Research, 19, 1978, cleaner, size-homogeneous and non-aggregated plate- 793-826.

101 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

4. H. J. Weiss, Platelet physiology and abnormalities of membranes, J. Clin. Invest., 45, 1966, 14-28. platelet function, N. Engl. J. Med., 293, 53 1, 1975, 9. R. L. Nachman, A. Hubbard, B. Ferris, Iodination of 580-588. the human platelet membrane. Studies on the major 5. I. Djaffar, D. Vilette, P. Bray. Quantitative isolation surface glycoprotein, J. Bid. Chm., 248, 1973, 2928- of RNA from human platelets, Thrombosis Research, 36. 62, 3, 1991, 127-135. 10. R.L. Nachman, E.A. Jaffe, Subcellular platelet fac- 6. A.J. Marcus, M.B. Zucker, The physiology of blood tor VI 11 antigen and von Willebrand factor, J. Exp. platelets: Recent biochemical, morphologic and clini- Med., 141, 1975, 1101-1113. cal research, Grune & Stratton, Inc., New York., 1965, 11. B. Walkowiak, U. Kralisz, L. Michalec, Compari- 13-24. son of platelet aggregability and P-selectin surface 7. S.J. Shattil, R.A. Cooper, Role of membrane lipid expression on platelets Isolated by different meth- composition, organization and fluidity in human ods, Thrombosis Research, 99, 5, 2000, 495-502. platelet function, in: Progress in Hemostasis and 12. K. Rübsaamen, Dissertation zur Erlangung des Thrombosis, 4, Ed. T.H. Spaet, Grune & Stratton, Doktorgrades der Naturwissenschaften der New York, 1978, 59-86. naturwissenschaftlichen Fakultät III - Biologie und 8. A.J. Marcus, D. Zucker-Franklin, L.B. Safier, H.L. Vorklinische Medizin der Universität Regensburg, Ullman, Studies on human platelet granules and 2010.

102 Journal of the University of ChemicalM. Mladenov, Technology J. Jordanov and Metallurgy, 47, 1, 2012, 103-108

PREPARATION OF SECONDARY STANDARDS FOR X-RAY FLUORESCENCE ANALYSIS IN THE PRODUCTION OF MOLYBDENUM CONCENTRATE

M. Mladenov1,2, J. Jordanov2

1University of Chemical Technology and Metallurgy Received 01 November 2011 8 Kl. Ohridski, 1756 Sofia, Bulgaria Accepted 16 January 2012 E-mail: [email protected] 2 Institute of General and Inorganic Chemistry of the Bulgarian Academy of Sciences, Department of Analytical Chemistry, X-ray fluorescence laboratory, 11 Acad. G. Bontchev Str., 1113 Sofia, Bulgaria

ABSTRACT

In practical applications, for X-ray fluorescence analysis (XRF) of molybdenum concentrates, there is not elaborated and certificated standards. Factories and producers of molybdenum, work with own secondary standards or with the methods of the “wet chemistry”. In this aspect for obtaining of correct data from the analysis it is very important the secondary standards to be prepared correctly and precisely. In this work data from the preparation of secondary standards for quick quantitative XRF analysis are discussed. The obtained results from the analysis of Mo, Cu, Fe and Re by methods of Wave dispersive X-ray fluorescence (WD- XRF) analysis and total X-ray fluorescence (TXRF) analysis, in the preparation of secondary standards are also presented. Keywords: secondary standards; molybdenum concentrate; wave dispersive X-ray fluorescence (WD-XRF) analysis; total X-ray fluorescence (TXRF) analysis; practical use.

INTRODUCTION Different calibration methods for quantitative XRF analysis can be employed but the most frequently In the production, the express method for used is the method of the calibration curve. In this determination of the element concentrations in method, the calibration curve is drawn from data for molybdenum concentrates is indispensable. One such secondary standards with known concentrations of the method is the X-ray fluorescence (XRF) analysis. This analyzed elements. Because of this, it is necessary to is the most often used technique for analysis of the ore produce secondary standards, adequate in matrix and and the final products in the metallurgical process. The phase composition to the ore deposit from which the method is fast and effective and it provides required ore is mined [4, 5]. accuracy and reproducibility, both for tracking of the The total X-ray fluorescence (TXRF) method is process and for the certification of the final products a practically new method which over the past several [1-3]. years has become an established analytical technique

103 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012 for the multi-element determination of trace elements of molybdenum concentration in the samples and total in various types of samples [6-10]. This method is X-ray fluorescence analyzer S2 PICOFOX (Bruker AXS, basically an energy dispersive XRF technique with a Germany), for determination of the concentrations of special geometry. The properties of the method together rhenium (Re), cuprum (Cu) and iron (Fe) were used. with recent instrumental improvements have been Laboratory muffle-furnace LM-212.11 type, with reported by many authors [11-13]. TXRF has been temperature range up to 1 500 oC and temperature extensively used not only for the elemental analysis of control within ±2 oC deviation was used for the pro- various materials but also for depth profiling and surface cess of melting. The samples and the synthetic stan- analysis. dards were melted in platinum-golden crucibles (95 In the current article, the experimental work for % Pt, 5 % Au) with a volume of 25 cm3. A hot plate preparation of the secondary standards is principal with possibility for regulation of the temperature with elaboration of the method described in [4]. In the ±2 oC, was used in appropriate thermal regime in the literature we don’t detect similar elaborations for effusion and the cool down of the melts. An ensemble preparation of secondary standards for XRF analysis of composed by an iron form and an aluminum stamp molybdenum concentrates. (Fig. 1), was employed in the effusion of the melt and The aim of the current work is the analytical it mould in disc form. tracking of the dressing process for molybdenum The following reagents and elements were used concentrates, through preparation and use of secondary for the experiments: CuO (B.D.H., pure, England), CaO standards for XRF analysis. Here, an important and Fe powder (Merck, p.a., Darmstadt, Germany), MoS 2 requirement is the accurately determination of the (Alfa Aesar, Karlsruhe, Germany), SiO (Chimsnab, 2 elements Mo, Cu, Fe and Re in these secondary pure, Dimitrovgrad, Bulgaria) and Triton X-100 (B.D.H., standards. England). Ga(NO ) .8H O (Johnson Matthey and Co, 3 3 2 London, England) was used for preparation of the EXPERIMENTAL internal standard for the TXRF analysis. The used melting reagents were: Spectromelt A10 Wave dispersive X-ray fluorescence spectrometer (Li B O ) (Merck, Darmstadt, Germany), LiCO 2 4 7 3 VRA 20 (Carl Zeiss Jena, Germany), for determination (Chemapol, Praha, Chechoslovakia), KNaCO (KCO / 3 3

1 5 m m

2 4 m m h = 52 mm

h = 6 mm

h = 8 mm

4 4 m m

3 9 mm

3 8 mm 1 mm

3 h = 8 mm mm

2 mm

5 0 mm

Fig. 1. Ensemble for effusion of the melts.

104 M. Mladenov, J. Jordanov

Table 1. Contents of the elements in the synthetic standards

Standard Ìî, % Cu, % Fe, % Ca, % Si, % Mi1 35 3 0.5 2 59.5 Mi2 40 1 3 2 55 Mi3 45 2 5 2 46 Mi4 50 1.5 2 2 44.5 Mi5 55 0.5 1 2 41.5

Na CO ) (Riedel de Haen, Hanover, Germany) and The used methods in details are given below: 2 3 NaNO (Fluka AG, pure, Buchs, Switzerland). Melting method 3 1. Twelve samples covering the interval of con- RESULTS AND DISCUSSION centrations for element Mo in the ore, are selected. Determination of Mo 2. By mixing of appropriate chemical compounds Determination of the Mo concentrations in the in right ratios, which cover the full interval of secondary standards samples was done by the method concentrations (see Table 1), the synthetic primary of XRF. In general, the following procedure is used: 1) standards are prepared. five primary standards for calibration of the XRF 3. The accurately weighed (±0.00001 g) amounts apparatus are prepared. These five standards are melted of the components are mixed for preparation of the and after forming of glass discs they are used for amounts of synthetic primary standards. calibration of the XRF apparatus and the calibration 4. By 0.2 g from each sample and each synthetic curve preparation; 2) twelve samples of the Mo-ore, standard, are mixed with respective quantities from the are mixed with the melting reagents and are melted. melting reagents in ratio 1:32 [4]. The used quantities These samples represent the secondary standards. After of the melting reagents are: Spectromelt A10 – 4.2 g; their cooling and obtaining of the glass discs, the LiCO – 1.6 g; KNaCO – 0.8 g; and NaNO – 0.5 g. 3 3 3 prepared calibration curve was used to determine con- 5. The mixtures are melted at temperature 1120 centration of Mo in the secondary standards. oC for 90 min, and are effused into the ensemble. By

55

50

45

Mo (%) Y = A + B * X Parameter Value 40 A -5,67939 B 0,00139 R = 0,99569 35

28000 30000 32000 34000 36000 38000 40000 42000 44000 CPS Fig. 2. Glass-mode disc of the synthetic standard Fig. 3. Calibration curve for Mo

105 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

Table 2. Contents of Mo, Fe, Cu and Re in the secondary standards.

Standard Ìî, % Fe, % Cu, % Re, ppm P1 18.55±0.15 4.05±0.02 4.01±0.03 1520±24 P2 38.31±0.12 1.52±0.13 1.30±0.03 1360±23 P3 22.65±0.40 4.57±0.15 4.52±0.04 1710±33 P4 38.08±0.56 2.73±0.10 2.03±0.03 1530±81 P5 - 0.64±0.10 0.33±0.04 1190±33 P6 44.59±0.10 0.68±0.06 0.40±0.01 1130±12 P7 49.51±0.55 0.48±0.02 0.25±0.01 1130±59 Ð8 - 0.49±0.04 0.25±0.01 1240±2 P9 37.07±0.49 2.77±0.30 2.54±0.25 1570±65 P10 49.15±0.38 0.77±0.01 0.45±0.01 1490±59 P11 37.56±0.36 1.70±0.01 1.29±0.01 1520±23 P12 43.98±0.28 1.11±0.09 0.57±0.07 1370±83 P13 - 1.26±0.22 0.82±0.10 1425±94 P14 42.28±0.33 2.25±0.25 1.60±0.01 2030±28 P15 44.39±0.45 1.47±0.01 1.18±0.01 1440±21

this process the phase structure and the granulometric with Rh anode, voltage – 35 kV, current – 45 mA, crys- composition of the samples are equalized, and also the tal – LiF , detector – SZ, collimator – 0.14 o, dis- 220 matrix influences are eliminated. criminator – 2-8 V and measuring time – 20 s. 6. The cooling of the melt has been done by Determination of Cu, Fe and Re reaching 120 oC and keeping for 20 min, and then For determination of the concentrations of the cooling down to room temperature, on a turned-off elements Cu, Fe and Re the method of TXRF was used. plate. 1. The fifteen samples are ground in an agate Table 1 presents the compositions of the synthetic mortar and from there samples of 0.04000 g are weighed standards. (on analytical balance with accuracy ±0.00001 g) into “Eppendorf” test-tubes with volume 2.0 ml. To that XRF method quantity 1.5 ml of detergent solution and 10 ìl from 1. The obtained glass-mode synthetic primary solution of the „internal standard” (1 mg/L Ga) are added standards (Fig. 2) are used for calibration of the XRF [14]. apparatus. 2. The prepared suspensions in portions of 10 ìl 2. By the analysis of the glass-mode synthetic are poured drop-wise upon quartz discs. After that the primary standards the standard curve for Mo is prepared discs are dried under vacuum. (Fig. 3). 3. The quartz discs with the fixed on their surface 3. The concentrations of Mo in the selected samples are putted into the apparatus for measuring. samples, which now are called “secondary standards”, Three parallel measurements are done for each sample. are determined from the obtained calibration curve. 4. The used measuring condition for the TXRF 4. The analysis is done under the following work- apparatus were: X-ray tube with Mo target, Ni/C mono- ing conditions: line – Ìo-Êα (2θ = 22.89o), X-ray tube chromator with 80 % reflection and a liquid nitrogen

106 M. Mladenov, J. Jordanov

free Silicon Drift Detector (SDD) with an energy reso- in metal sludge/waste oil derived adsorbents: Effect lution of < 159 eV, voltage – 50 kV, current – 1 A and of heat treatment, Chemical Engineering Journal, measuring time – 300 s. 138, 2008, 155–165. 3. G. Rossini, A. M. Bernardes, Galvanic sludge metals RESULTS recovery by pyrometallurgical and hydrometallurgical treatment, Journal of Hazardous Materials, 131, The determined concentrations of the elements 2006, 210–216. in the secondary standards are shown in Table 2. 4. L. Vuchkova, J. Jordanov, Fusion method for The obtained results for the element Mo in the preparation of refractory nickel based alloy powders secondary standards cover the concentration interval for X-ray fluorescence spectrometry, The Analyst, from 18 to 50 % and allow for analysis of the 125, 2000, 1681-1685. molybdenum concentrate from the beginning of the 5. J. Jordanov, L. Vuchkova, A. Mitev, Preparation of process to its end (after enriching of the ore). reference samples for X-ray fluorescence analysis of The determined concentrations of the elements cement flours, Analytical Laboratory, 7(4), 1998, Cu, Fe and Re in the secondary standards, allow their 235-236. content to be followed in the ore during the technological 6. P. Hoffmann, V. K. Karandashev, T. Sinner, H. M. process. These results can be used also for taking Ortner, Chemical analysis of rain and snow samples decisions on the final products, for example: reuse, from Chernogolovka/Russia by IC, TXRF and ICP- certification, sale, etc. MS, Fresenius J. Anal. Chem., 357, 1997, 1142-1148. The prepared secondary standards were applied 7. H. Fink, U. Panne, M. Theisen, R. Niessner, T. Probst, in factory for production of Mo concentrate in the X. Lin, Determination of metal additives and Republic of Bulgaria. Before their application, the results bromine in recycled thermoplasts from electronic from analysis for Mo were obtained after 1-2 days, by waste by TXRF analysis, Fresenius J. Anal. Chem., chemical method and after that – within 15 min. The 368, 2000, 235–239. prepared secondary standards now are successfully used 8. I. Varga, Á. Szebeni, N. Szoboszlai, B. Kovács, De- in the practical process for the molybdenum concentrates termination of trace elements in human liver biopsy production, for fast, effective and adequately accurate samples by ICP–MS and TXRF: hepatic steatosis control of the concentrations of the above elements. and nickel accumulation, Anal. Bioanal. Chem., 383, 2005, 476–482. CONCLUSIONS 9. B. U. Peschel, U. E. A. Fittschen, G. Pepponi, C. Jokubonis, Ch. Streli, P. Wobrauschek, G. Fifteen secondary standards for the production Falkenberg, J. A. C. Broekaert, Direct analysis of process of Mo-concentrates were prepared. In twelve of Al O powders by total reflection X-ray fluorescence 2 3 them the Mo concentration and in all of them the spectrometry, Anal. Bioanal. Chem., 382, 2005, concentrations of Cu, Fe and Re were determined. 1958–1964. After application of the secondary standards the 10. C. Vázquez, M. S. Maier, S. D. Parera, H. Yacobaccio, time for analysis of Mo in the ore was very significantly P. Solá, Combining, TXRF, FT-IR and GC–MS infor- shortened. mation for identification of inorganic and organic components in black pigments of rock art from Alero REFERENCES Hornillos 2 (Jujuy, Argentina), Anal. Bioanal. Chem., 391, 2008, 1381–1387. 1. V. Šprta, B. Knob, P. Janoš, X-ray fluorescence determi- 11. R. S. Hockett, Advances in X-ray Analysis, An update nation of total sulfur in fly ash, Fresenius J. Anal. on standards activity for TXRF and the challenges ahead, Chem., 364, 1999, 705–708. Advances in X-ray Analysis, 39, 1997, 767-770. 2. K. Kante, J. Qiu, Z. Zhao, Yu Cheng, T. J. Bandosz, 12. M. Dargie, A. Markowicz, A. Tajani, V. Valkovic, Development of surface porosity and catalytic activity Optimized sample preparation procedures for the

107 Journal of the University of Chemical Technology and Metallurgy, 47, 1, 2012

analysis of solid materials by total-reflection XRF, 1997, 243-245. Fresenius J. Anal. Chem., 357, 1997, 589-593. 14. Ar. Gross, H. Stosnach, Lab Report XRF 438 – S2 13. Y. H. Tian, K. Liu, X. R. Wu, R. G. Wang, A small PICOFOX: Ultratrace element analysis of TXRF analyzer and its applications, Journal of nanoparticles, Bruker AXS, Germany, 2009 – Radioanalytical and Nuclear Chemistry, 217, 2, available online.

108 Journal of the University of Chemical Technology and Metallurgy

GUIDE FOR AUTHORS

The Journal of the University of Chemical Technol- Submission of a manuscript implies it is not being simul- ogy and Metallurgy, a specialised scientific edition, pub- taneously considered for publication elsewhere and that the lishes original research in the field of chemical tech- authors have obtained the necessary authority for publication. nology and metallurgy, chemical engineering, biotech- All papers will be independently refereed. nology, industrial automation, environmental protec- Language tion and natural sciences. The following topics are used All papers submitted to the journal should be written to identify articles published in the Journal of the Uni- in good English. Authors are encouraged to have their pa- versity of Chemical Technology and Metallurgy: per checked before submission for grammar and clarity. Organic Synthesis and Fuels; Polymer Engineering; Manuscript preparation Textile and Leather; Cellulose, Paper and Polygraphy; General: Two copies (for Bulgarians - two in En- Inorganic and Electrochemical Productions; glish and one in Bulgarian) should be provided in double- Physical Metallurgy; Metallurgy of Iron and Cast- space typing on numbered pages of uniform size. A font ing; Metallurgy of Non-ferrous and Semiconducting size of 12 pt is required. The name, complete postal ad- Compounds; Technology of Silicates; dress, telephone number, fax number and e-mail address Chemical Engineering; Automation of Industry; of the corresponding author should be given on the first Information Technology; Biotechnology; Economics and page of the manuscript. Full postal addresses must be Management of Industry; Sustainable Development and given for all co-authors. The Editor reserves the right to Environmental Protection; adjust style to certain standards of uniformity. Physics; Applied Mechanics; Programming and Paper length: Research papers should not exceed Computer Systems; Electrotechnics and Electronics; 10 pages (8000 words). Communications must not ex- Inorganic Chemistry; Organic Chemistry; Analyti- ceed 4 printed pages and will be given priority for rapid cal Chemistry; Physical Chemistry; publication. Word counts include text, references, fig- Management of the Education Quality. ures and tables. The number of figures and/or tables The Journal of the University of Chemical Technol- should not exceed eight. Editor may allow longer pa- ogy and Metallurgy publishes regular articles, reviews pers of significance and originality. Every manuscript (by invitation only), short communications, letters to should include: Abstract, Introduction, Experimental, the editor, and book reviews. The main distinction be- Results and Discussion, Conclusions, References. tween regular articles and short communications is that Abstract: The Abstract summarizes the main find- the latter may be shorter. Letters to the editor provide ings of the paper in 100 to 250 words. After the ab- accelerated publication of very brief accounts of experi- stract a list of up to 6 keywords that will be useful for mental findings, theoretical developments, or comments indexing or searching should be included. on previously published work. The Introduction should be as concise as possible, The experimental, theoretical and numerical pro- without subheadings. Special attention must be paid on cedures must be clearly described. the latest literature sources. Submission of Manuscripts The Experimental section should be sufficiently Papers should be submitted to the Editor-in-Chief detailed to enable the experiments to be reproduced. Bogdana Koumanova at the following address: A reference to the widely used standard methods should University of Chemical Technology and Metallurgy be given. Methods in detail should be given only if 8 Kliment Ohridski Blvd., Sofia 1756, Bulgaria, (JUCTM) they represent a new approach. Manufacturers’ names Authors could submit their papers also electroni- and model numbers for the equipment used are usu- cally to e-mail: [email protected]. ally desirable.

109 Journal of the University of Chemical Technology and Metallurgy

Results and Discussion may be organized into sub- Units should be given in SI system. They should be headings. Figures and tables should not be included into in the form, e.g. g cm-1 rather than g/cm. the text. The notes should be given at their suitable places. Figures should be numbered consecutively with Acknowledgements when needed should be brief Arabic numerals. All figures should be supplied on and should precede the references. separate sheets. All illustrations must be readable References should be cited in the text by an Arabic when reduced to a width of 75 mm (single column numeral in square brackets, listed numerically and typed figure) or 160 mm (double column figure). Photo- double-spaced on a separate sheet at the end of the pa- graphs, charts and diagrams are all to be referred to per. References should be arranged in the order in which as “Figure(s)” and should be numbered consecutively they appear in the text. Journal titles should be abbre- in the order to which they are referred. They should viated according to the latest Chemical Abstracts Ser- accompany the manuscript, but should not be included vice Source Index. Only articles that have been pub- lished or are in press should be included in the refer- within the text. All illustrations should be clearly ences. Unpublished results should be cited as such in marked with the figure number. Figure captions the text. References in non-English language should be should be typed double-spaced on a separate sheet. translated in English. The original language must be Figures should not include text. Lines should be black, shown in brackets. In the reference list, the styling, punc- of an adequate thickness (around 1 pt) and curves tuation and capitalization should be as follows: should be smooth. Because of technical complica- For journals: tions that can arise in converting colour figures to 1. D.R. Mehandjiev, P.R. Bontchev, Application of “gray scale”, please submit black-and-white files. magnetic methods to the study of 3d transition metal Graphic files should be in TIFF or JPG with resolu- complexes, J. Univ. Chem.Technol. Met. (Sofia), 45, 4, 2010, 415-420. tion not less than 300 dpi. For books: Tables should be numbered consecutively with 2. A. Krueger, Carbon Materials and Nanotechnology, Arabic numerals in the order of appearance in the text. Weinheim, WILEY-VCH Verlag CmbH & Co. KGaA, Each table should be given on separate sheet with a 2010. short descriptive title directly above it, with essential For edited books: footnotes below. 3. F.W. Cooke, J.E. Lemons, B.D. Ratner, in: B.D. Ratner, Two identical copies of the manuscript in English A.S. Hoffman, F.J. Schoen, J.E. Lemons (Eds.), Biomaterials and one in Bulgarian (for Bulgarian authors) with a set Science, Academic Press, San Diego, 1996, p. 11. of original illustrations and a disk must be submitted. For conference proceedings, symposia, etc.: The file with the manuscript should be written in 4. E. Motta, D. Rajpathak, Z. Zdrahal, R. Roy, The Epistemology of Scheduling Problems, F. V. Harmelen, double-density or High density 3.3" disk formatted for editor, Proceedings of the 15th European Conference IBM PC compatibles. The paper should be prepared on Artificial Intelligence, Lyon, France, 2002, 215-219. by word processor MS Word 6.00 or higher with true- Formulae should be typewritten (Equation Editor) with type font Times New Roman size 12 with double line ample space around them. The meanings of all symbols spacing. The disk format, word processor format, file have to be given immediately after the equation in which names, the title and the authors must be indicated on they are first used. Equations should be sequentially num- the disk. Experts advising the Editor will review all pa- bered (on the right side of the equation and in parenthe- pers. Authors make the revision of proofs before print- ses). Subscripts and superscripts should be clearly indi- ing. They will receive the electronic form of the paper. cated. Greek letters and other non-Latin symbols should Manuscripts of papers submitted to the journal will not be typed using Times New Roman Gr and Symbols. be sent back to the authors.

110