887THE DEPARTMENT OF CHEMISTRY, ETH ZRICH, IN THE NEW HNGGERBERG BUILDINGSCHIMIA 2001, 55, No. 10

Chimia 55 (2001) 887891 Schweizerische Chemische Gesellschaft

ISSN 00094293

Environmentally Oriented Designand Assessment of Chemical Productsand Processes

Maximilian Stroebe, Volker H. Hoffmann, Andreas Zogg, Martin Scheringer, and KonradHungerbhler*

Abstract: In this paper we illustrate our work on integrated process and product design and assessment usingthree examples: a small calorimeter for quick reaction screening, a multi-criteria process technologyassessment method, and a screening tool for exposure-based risk assessment of chemical products. A shortintroduction is given on our general framework for considering environmental objectives in chemical productand process design.

Keywords: Calorimetry Environmental product design Exposure-based assessment Multi-criteria decision-making Sustainability

The specialty chemicals industry today isfacing the challenge of succeeding withits products and processes in a world ofcontinuously growing economic and eco-logical requirements. We at the group forEnvironmental and Safety Technologyin Chemistry focus our teaching and re-search on delivering instruments to im-prove decision making that considersboth economic and environmental objec-tives in early design stages (Hungerbh-ler et al. [1]). This proactive way of envi-ronmentally oriented product and processdevelopment is an iterative procedurecomprising the generation of design al-ternatives, performing mass and energybalances for these technology options,analysing the benefits and risks usingeconomic value and ecological impactdata, and finally evaluating the results bycomparing them with value-based devel-

*Correspondence: Prof. Dr. K. HungerbhlerDepartment of ChemistrySwiss Federal Institute of TechnologyETH HnggerbergCH8093 ZrichTel.: + 41 1 632 60 98Fax: + 41 1 632 11 89E-Mail: hungerb@tech.chem.ethz.ch

opment aims and protection goals (Fig. 1).According to our problem understanding,eco-efficiency, inherent safety and socialacceptance are the key principles of thisdevelopment approach. Special attentionhas to be given to the problem of model

and data uncertainty. We also implementthese concepts in our teaching activitiesusing interdisciplinary, teamwork-orient-ed case studies developed in cooperationwith the chemical industry (Fenner et al.[2]).

Fig. 1. Framework for integrated process and product design.

anal

ysis

values, indicators

assessment

effe

ct d

ata,

scen

ario

s

tech

nic

al s

yste

mim

pact

benefits & risks

design guidelines:

eco-efficiency

inherent safety

social acceptability

alte

rnat

ives

system data

technicalsystem

desi

gn

modelling

THE DEPARTMENT OF CHEMISTRY, ETH ZRICH, IN THE NEW HNGGERBERG BUILDINGS 888CHIMIA 2001, 55, No. 10

This combination of two calorimetricmeasurement principles allows for thedirect measurement of the heat evolvedduring a reaction without a mathematicalbaseline correction. This new reactorconcept and its realization with Peltier el-ements and a Teflon-coated metal blockhas been patented and is currently underconsideration for commercial use. Thecalorimeter can be operated at tempera-tures between 20 and 200 C and a pres-sure of 0 up to 106 Pa. It is equipped witha magnetic coupling for the stirrer, withpressure and temperature sensors and twoinlets and is combined with an FT-IRspectrometer.

For processing the measured data, notools are currently available for a simul-taneous evaluation of both the thermaland the spectroscopic signal (examplessee the Scheme and Fig. 3). Therefore weare developing an evaluation algorithmusing non-linear optimisation to estimateunknown reaction parameters such as theheat of reaction, the rate constants, andthe phase transfer coefficients.

Fig. 2. Small reaction calorimeter with integrat-ed FT-IR probe. Fig. 4. Common HCN synthesis reactions.

Fig. 3. Heat of the reaction versus time course (left) and IR spectra measured online (right).

Scheme. The heterogeneous reaction investigated.

In this paper, three topics of our re-search are presented: (1) a small calori-meter combined with an FT-IR spectro-meter for quick reaction screening, (2) amulti-criteria process technology assess-ment tool and (3) a screening tool for theexposure-based assessment of chemicals.

1. Reaction Screening Using aSmall Calorimeter Combined with aFT-IR Spectrometer

Optimised reaction conditions implya better environmental performance of achemical process because raw materialutilisation is improved, the energy de-mand for separations is decreased, andthe amount of waste is reduced. To meetthis goal, a systematic and quick determi-nation of kinetic and thermodynamic re-action parameters during early processdevelopment is needed. For this purpose,a small reaction calorimeter with inte-grated FT-IR probe has been developed.Such a device is of particular importancefor new fine-chemical and pharmaceuti-cal products where generally only smallamounts of test substance are availableand time-to-market is crucial (Pastre etal. [3]).

The reaction calorimeter (Fig. 2) has asample volume of 2045 ml. To measurethe thermokinetic parameters at isother-mal conditions, the power compensationprinciple (internal compensation heating)has been combined with the heat balanceapproach (external Peltier elements).

2. Multi-Criteria ProcessTechnology Screening

Multi-criteria process assessmentdeals with the question of how to evalu-ate and compare different process tech-nologies in a systematic way. To developand illustrate the methodology, the well-known HCN production technology witha large number of process alternativeswas chosen (Hoffmann [4]). HCN ismostly produced from a carbon feedstockand ammonia according to the reactionsshown in Fig. 4.

A general process structure with dif-ferent system boundaries is indicated inFig. 5. While a number of standard tech-nologies are currently used in industry,the systematic combination of the litera-ture-known unit operation technologiesbuilding up the process superstructuregenerates more than 70'000 process alter-natives (for an example see Fig. 6). Byintroducing additional constraints such ascompatibility between the different unitoperation technologies, safety require-

889THE DEPARTMENT OF CHEMISTRY, ETH ZRICH, IN THE NEW HNGGERBERG BUILDINGSCHIMIA 2001, 55, No. 10

ments or emission limits, this number canbe reduced to about 1200 feasible processoptions. Using a flow-sheeting program,mass and energy balances for each ofthese alternatives are calculated and aresubsequently evaluated using economicand ecological indicators. Fig. 7 shows atwo-dimensional plot (Pareto Plot) inwhich the Total Annualised Profit PerService unit, TAPPS, and the MaterialIntensity Per Service unit, MIPS, havebeen chosen to evaluate the differenttechnologies. The TAPPS is calculated asthe sum of sales, raw material costs andannualised investment cost per kilogramof main product, while the MIPS assumesa life-cycle perspective and adds the ma-terial use of all upstream products that arenecessary to produce one kilogram ofmain product. Hence, in Fig. 7 a processin the upper left corner characterized bya high profit and a low material intensity is preferable to alternatives to the lowerright.

Fig. 7 illustrates that alternatives tendto form clusters depending on the key de-cision of which reactor technology ischosen. Generally, BMA processes seemto exhibit a better performance than And-russow processes, while Fluohmic proc-esses are environmentally significantlyworse than the other two technologiesdue to the extensive use of electricity tooperate the reactor. In other words, if aninferior basic choice is made, e.g. to con-struct a Fluohmic reactor, all other proc-ess decisions cannot improve the processin such a way that it becomes comparableto the performance of an Andrussow oreven a BMA process. Once the reactor ischosen, a more detailed analysis showsthat a second key decision concerns theuse of the by-product hydrogen. This

suggests that, for technology screening, aproper analysis of potential uses for hy-drogen has a higher priority than, for ex-ample, the design of the purification sys-tem. In order to direct research and devel-opment efforts and to allocate financialresources to the most promising process-es, key decisions of this type need to beidentified and inferior processes have tobe excluded as early as possible (Heinzleet al. [5]).

Unfortunately, it is an inherent char-acteristic of technology assessment inearly development stages that decisions

Fig. 5. Block diagram for HCN production with different system boundaries.

Fig. 6. Unit operations network for HCN production.

Fig. 7. Pareto plot of the environmental andeconomic assessment of 1244 HCN produc-tion technologies.

THE DEPARTMENT OF CHEMISTRY, ETH ZRICH, IN THE NEW HNGGERBERG BUILDINGS 890CHIMIA 2001, 55, No. 10

have to be made with incomplete infor-mation. A probabilistic approach basedon Monte-Carlo simulations has been de-veloped allowing to efficiently propagateparameter uncertainties through complexprocess models and to identify key uncer-tainties (Hoffmann [4]). While the abso-lute uncertainty intervals of differenttechnologies might partly overlap, a rela-tive analysis can enable the decisionmaker to make a ranking according to theeconomic and environmental perform-ance.

3. A Screening Tool for Exposure-Based Assessment of Chemicals

In the design of chemicals, the assess-ment of the environmental fate and toxic-ity is required to ensure a good environ-mental performance. However, a completeunderstanding of a chemicals behaviourin the environment including its eco-tox-icity requires tremendous efforts (testingfor degradability, bioaccumulation, andtoxicity to different species and ecosys-tems). For this reason, it is helpful to as-sess the environmental fate (chemical re-activity; transport with wind and water

currents; exchange between the environ-mental media) and the eco-toxicity of achemical in two subsequent steps. Espe-cially on a screening level, when toxicitydata are scarce, it is appropriate to inves-tigate the exposure potential of a chemi-cal as a clearly defined first assessmentstep (exposure-based assessment). If achemical exhibits an unwanted exposurepotential (high persistence and mobilitysuch as chlorofluorocarbons), it can beconcluded - even without knowledge ofits toxicity - that the chemical is likely tocause environmental problems.

Measures of the exposure potential ofa chemical are its persistence, , and spa-tial range, R, which describe the temporaland spatial extent of environmental expo-sure patterns (Fig. 8). For screening pur-poses, it is most useful to determine Rand with generic multimedia fate mod-els (Van de Meent et al. [6]). The Chem-range model (Scheringer et al. [7]) is asimple model of the global circulationthat calculates R and for organic chemi-cals from their Henrys law constants,octanol-water partitioning coefficientsand average degradation rates in soil, wa-ter, and air. The model results can be usedfor the ranking and classification of

chemicals. Fig. 9 shows model results fordifferent substances, some of which arePersistent Organic Pollutants (POPs)with high uncertainty in their atmospher-ic degradability and, therefore, also in thespatial range. Within the possibilities of ageneric model, the model results agreequite well with experimental findings onthe global dispersion of chemicals (Sche-ringer [8]).

Recently, Fenner et al. [9] developeda methodology to include transformationproducts together with the parent com-pound into the exposure-based assess-ment. To this end, the joint persistencecovering the entire cascade of the parentcompound and a selection of relevantmetabolites has been introduced. Appli-cation to different chemicals such asatrazine, perchloroethylene, methyl-t-butylether, and nonylphenol ethoxylatesrevealed that the inclusion of transforma-tion products can cause considerablechanges to the ranking of chemicals ac-cording to their persistence.

The approach of exposure-based as-sessment introduces a new perspectiveinto chemicals assessment. In contrast totoxicity indicators, which do not provideany information on the location and time

Fig. 8. Calculation of the environmental per-sistence of a chemical [A] from the functionM(t) describing the decrease of the chemicalsoverall mass in time and calculation of itsenvironmental spatial range R [B] as the 95%interquantile distance of the concentrationdistribution. Reprinted with permission fromM. Scheringer, Env. Sci. Technol. 1996, 30,16521659. Copyright (1996) American Chem-ical Society.

891THE DEPARTMENT OF CHEMISTRY, ETH ZRICH, IN THE NEW HNGGERBERG BUILDINGSCHIMIA 2001, 55, No. 10

Fig. 9. Spatial range in air, R, vs. overall persistence t as calculated with the Chemrange model.All chemicals are released to the soil. Persistent Organic Pollutants exhibit high uncertainties inR due to their unknown atmospheric degradability. Reprinted with permission from R.L. Lipnicket al. (eds.) Persistent, Bioaccumulative, and Toxic Chemicals II, American Chemical Society2001, p. 59. Copyright (2001) American Chemical Society.

of the occurrence of effects, it explicitlyfocuses on the temporal and spatial distri-bution of benefits and burdens associatedwith the production and use of chemicals.If, in the future, exposure patterns couldbe restricted to the surroundings of theusers of a chemical, it is more likely thatthe users and those who are exposed tothe chemical are in similar political, eco-nomic, and legal contexts, which wouldmake negotiations much easier than onthe global level. In this sense, we suggestthat the chemical product technologyshould be directed towards a chemistryof low ranges.

AcknowledgementsWe thank Kathrin Fenner and Ulrich Fischer

for valuable discussions.

Received: June 21, 2001

[1] K. Hungerbhler, J. Ranke, T. Mettier,Chemische Produkte und Prozesse. Grund-konzepte zum umweltorientierten Design,Springer, Berlin, 1998.

[2] K. Fenner, G. Jdicke, P. Alean-Kirk-

patrick, K. Hungerbhler, Chemosphere2001, 43, 5-16.

[3] J. Pastre, A. Zogg, U. Fischer, K. Hunger-bhler, Organic Process Research & Devel-opment 2001, 5, 158166.

[4] V. Hoffmann, Multi-Objective DecisionMaking Under Uncertainty in ChemicalProcess Design, Dissertation 14140, De-partment of Chemistry, ETH Zrich, 2001.

[5] E. Heinzle, D. Weirich, F. Brogli, V. Hoff-mann, G. Koller, M. A. Verduyn, K. Hun-gerbhler, Ind. Eng. Chem. Res. 1998, 37,33953407.

[6] D. Van de Meent, T. E. McKone, T. Parker-ton, M. Matthies, M. Scheringer, F. Wania,R. Purdy, D. H. Bennett, Persistence andTransport Potential of Chemicals in a Multi-media Environment, in Evaluation of Per-sistence and Long-Range Transport of Or-ganic Chemicals in the Environment, Eds.G. Klecka et al., SETAC Press, Pensacola,2001, p. 169204.

[7] M. Scheringer, H. Held, M. Stroebe, Chem-range 1.0 - A Multimedia Transport Modelfor Calculating Persistence and SpatialRange of Organic Chemicals, Laboratoryof Chemical Engineering, ETH Zrich,http://ltcmail.ethz.ch/hungerb/research/product/chemrange.html, 2001.

[8] M. Scheringer, Environmental Science &Technology 1997, 31, 10, 28912897.

[9] K. Fenner, M. Scheringer, K. Hungerbhler,Environmental Science & Technology 2000,34, 17, 38093817.