A Simulator For Explaining Organic Reactions Through

Qualitative Reasoning

Introduction (I)

• We implemented qualitative reasoning based on QPT ontology in a software tool (QRiOM)

• The main educational goals of the tool are that:– The students will undergo mental change so that they are

able to explain chemical phenomenon in a more elaborated way

– The students would benefit from it in terms of improving their reasoning skills and enhancing their understanding of the organic processes

Introduction (II)• The main focus of the design of QRiOM is on

the generation of explanation (causal and behavioral aspects of organic reactions)

• The results of a simulation is presented in five main forms:– Names of the final product (most stable form), and organic

mechanism used to predict the output– QPT processes (for inspection)– Causal diagram for tracing parameter dependency during a

make-bond/break-bond organic process– The entire reaction route (from initial substrate to the final

products)– Reacting species (view pairs) used in each reaction step and

their chemical parameters’ states change

Inputs 1 (Substrates &

Reagents)

2 Substrates Recognizer

4 QPT model constructor

3 Chemical

Knowledge Base 1. Chemical facts 2. Chemical theories 3. OntoRM 4. QPT process models

5 Qualitative Simulator

1. Quantity space analyzer 2. Molecule update routine

6 Explanation Generator (Causal and behavioral)

7

Outputs 1. Final products 2. Reaction routes 3. mechanism used 4. explanation notes

This module checks user selection, and returns the “type”

of the inputs as either a nucleophile or an electrophile. From here, an organic reaction

may be determined

This module will automate the construction of a

process model using QPT modeling constructs

The actual reasoning and simulation starts here. The task

is handled by several sub-modules. These include QSA

and MUR

This module will generate explanation to justify the

simulated result

Functional components of QRiOM

Functional components of QRiOM

Main interface of QRiOM simulator prototype

A B C

More learning activities and explanation can be viewed by clicking A, B and C buttons

Input selection Predicted products

Learners may inspect the automated

models

QPT processes

QPT model inspection page

Brief explanation of QPT slots is provided

Check how twoparameters are relatedin a given bond activity

The series of organic reactionsoccurred are stored in special purpose

data structures during reasoning in orderto generate causal graph;

like this one

Causal diagram inspection page

This button will display thepop-up window to tell the user how to read the causal diagram

More snapshots are provided for learners to examine the chemical parameter states’

changes

The entire reaction route is displayed after a simulation is completed

Basic properties of each atom

involved in a reaction are

recorded for user inspection

More learning activities and explanation

Learners can select any reacting species to study their

parameter history (from the first reaction step until the entire simulation is ended)

During simulation each chemical state change is recorded

Contents of the View Structure give the pairs of reacting species used in each small reaction step

Editor page for adding/deleting chemical facts and theories

Some organic chemistry terms are also provided; the page can be accessed from the main interface

Examples used are:Reasoning casesOrganic Mechanisms = 2 (SN1 and SN2)General Reaction formulas = 3

SN1Tertiary alcohol + Hydrogen halide(CH3)3COH HX

• CH3CH3CH3COH + HF• CH3CH3CH3COH + HCl• CH3CH3CH3COH + HBr• CH3CH3CH3COH + HI• CH3CH3CHOH + HF• CH3CH3CHOH + HCl• CH3CH3CHOH + HBr• CH3CH3CHOH + HI

Alkyl halide (tertiary) + Water molecules(CH3)3CX 2H2O (in excess)

• (CH3)3CF + 2H2O• (CH3)3CCl + 2H2O• (CH3)3CBr + 2H2O• (CH3)3CI + 2H2O

SN2

Alkyl Halide (primary) + incoming nucleophileCH3CH2X Hydroxyl functional group

CH3F + HO-CH3Cl + HO-CH3Br + HO-CH3I + HO-CH3CH2F + HO-CH3CH2Cl + HO-CH3CH2Br + HO-CH3CH2I + HO-CH3CH2CH2F + HO-CH3CH2CH2Cl + HO-CH3CH2CH2Br + HO-CH3CH2CH2I + HO-CH3CH2CH2CH2F + HO-CH3CH2CH2CH2Cl + HO-CH3CH2CH2CH2Br + HO-CH3CH2CH2CH2I + HO-

Other reaction formulas tested are:

1. CH3CH3CH3COH + HO-2. CH3CH3CH3COH + H2O 3. CH3OH + HO-4. CH3OH + H2O5. CH3CH3CHOH + HO-6. CH3CH3CHOH + H2O7. CH3CH2X + H2O8. CH3X + H2O9. CH3CH2X + X-10. CH3X + X-11. CH3CH2CH2CH2X + H2O12. CH3CH2CH2CH2X + HO-13. CH3CH2CH2X + H2O14. CH3CH2CH2X + HO-15. (CH3)3CX + HO-16. (CH3)3CX + X-

Conclusion (I)

• The simulation has been successfully tested with positive results– The QR approach enables prediction to be made, as

well as causal explanation generation about theories of a number of organic chemistry phenomena

– Cause-effect chain can be explained by using only the ontological primitives of QPT

• A study of learners’ feedback using the software was carried out– Overall, there was general understanding that the new

means of learning through qualitative simulation and explanation had proved valuable

Conclusions (II)

• The software – can predict final products– can explain its reasoning

• Unlike other chemistry software, QRiOM is implemented without having any pre-coded solution path in the knowledge base

• From a learner’s point of view– conceptual understanding is nurtured– reasoning ability is improved

Alicia Tang (University of Tenaga Nasional, MALAYSIA)Email: aliciat@uniten.edu.my Tel: +603 89212336

Sharifuddin Mohd. Zain (Dept. of Chemistry, Malaya University)

Noorsaadah Abd. Rahman (Dept. of Chemistry, Malaya University)

Rukaini Abdullah (Dept. of AI, Malaya University)