Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations KootK on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

Looking for an AI that will help me create a chemical functional group database

n8cole

Chemical
Jun 24, 2024
29
I am a chemical engineer working on a spreadsheet that calculates heat capacity using the Joback method (see links below if you need more information on the Joback method). For this calculator, I am creating a database of chemical compounds and their functional groups, but it is taking too long. I tried using AI like Claude to help speed up the process, but it proved to be too unreliable. Is there another AI or resource I could use to quickly generate the functional groups of chemical compounds? My goal is to be able to feed the AI or resource a list of chemical compounds and have it output the functional groups in a format that is easy to copy and paste into my spreadsheet. I expect this might not be possible, and I may have to go back to the slow and painful process of doing it manually, but I figured it was worth a shot.

I have attached an example of how my spreadsheet database is going to look.

Also, I don't know if I am asking the right group.

- Joback Online Calculator: https://www.egichem.com/tools/calculators/joback/
- Jaback wiki page: https://en.wikipedia.org/wiki/Jobac...oback method is a,the sum of group parameters.
 

Attachments

  • joback sample.xlsx
    16.5 KB · Views: 3
Replies continue below

Recommended for you

This almost looks like a textual analysis of the chemical's formula. If so you just need a parser not an AI.
 
AI is a generalized term for a lot of things. Claude is a family of large language models (LLMs). LLMs are given a large number of examples from which they find patterns based on proximity. If one gave an LLM only science fiction books and then asks how to get to the Moon it might suggest using a Star Trek transporter.

For one to help you there has to be an existing pattern to find and it has to have sufficient examples in order to find it. Generally this number is in the hundreds of thousands to tens of millions. You need one trained on the results you are looking for.
 
chatgpt please write me a script to break the chemical formula for an organic compound into groups as in the joback method

Python:
import re
from collections import defaultdict

def joback_method(formula):
    # Define functional groups and their corresponding atoms
    functional_groups = {
        'alkane': re.compile(r'(?P<C>\d+)C'),
        'alkene': re.compile(r'(?P<C>\d+)C(?P<H>\d+)H'),
        'alkyne': re.compile(r'(?P<C>\d+)C(?P<H>\d+)H'),
        'alcohol': re.compile(r'(?P<R>\d+)C(?P<H>\d+)OH'),
        'carboxylic acid': re.compile(r'(?P<R>\d+)C(?P<H>\d+)COOH'),
        'ketone': re.compile(r'(?P<R1>\d+)C(?P<H>\d+)C(?P<R2>\d+)O'),
        'aldehyde': re.compile(r'(?P<R>\d+)C(?P<H>\d+)CHO'),
        # Add more functional groups as needed
    }

    # Initialize a dictionary to count groups
    group_count = defaultdict(int)

    # Remove whitespace and split the formula into parts
    tokens = re.findall(r'([A-Z][a-z]?)(\d*)', formula)

    # Process each token
    for element, count in tokens:
        count = int(count) if count else 1
        if element == 'C':
            group_count['C'] += count
        elif element == 'H':
            group_count['H'] += count
        elif element == 'O':
            group_count['O'] += count
        # Add other elements if necessary

    # Analyze the groups based on Joback method
    for group, pattern in functional_groups.items():
        match = pattern.match(formula)
        if match:
            group_count[group] += 1

    return dict(group_count)

# Example usage
chemical_formula = "C3H8O"  # Example compound
result = joback_method(chemical_formula)
print("Functional Groups Count:", result)
 
Goodness knows if that is any help! I have my doubts.
 
Last edited:
Try Google's NotebookLM. https://notebooklm.google.com/

It works with the information sources that you feed it, so if you feed it the right stuff and ask it the right questions it might give you something useful.
 
Hi,
If I was in your shoes, I will ask my company to give me access to a database and/or ask them to purchase Yaws' Handbook of Thermodynamic and Physical Properties of Chemical Compounds - Yaws 2003. with thousands of components (Organic and Inorganic).
My view only.
Pierre
 

Part and Inventory Search

Sponsor