Register to reply

Using UV-Vis Spectroscopy to determine concentration

Share this thread:
FaNgS
#1
Mar26-12, 10:51 AM
P: 92
I have been conducting liquid phase adsorption experiments of xylene and toluene using activated carbon. I made solutions at 2 concentrations (200ppm and 400ppm) of individual components of toluene and xylene in methanol, as a solvent.

Initially, the plan was to use Gas Chromatography to determine the concentrations of the solutions to see how much of the toluene and xylene were adsorbed with time. After several weeks of struggling and calibrating the GC, I was not able to get good/repeatable results. The values output were not consistent and were varying up to 30% using standard solutions!

So I'm switching to UV-Vis spectroscopy but I have some doubts which I would like to clear. The plastic cuvette's that fit in the machine can hold up to about 1.5mL of liquid, but since I've only withdrawn 1mL of my adsorption experimental samples (at different times) will it be possible to dilute it and get results on the UV-Vis?
Is it correct to take 100μL of my sample and diluting it with 900μL of Methanol then multiplying by a factor of 9 to obtain the actual concentration?

These are the steps I will follow when using the UV-Vis:
1) I will fill a cuvette with methanol only and use that as a blank
2) I will calibrate with standard solutions of toluene and xylene of known concentrations of 200ppm and 400ppm.
3) Use my adsorption samples (diluting as mentioned above)

Are there any equations which I should use to determine the concentrations based on the components I have or can I use the Beer-Lambert Law?

Also are there any other methods I can investigate that can be used to determine the concentrations of my samples?
Phys.Org News Partner Chemistry news on Phys.org
Novel 'butterfly' molecule could build new sensors, photoenergy conversion devices
Rubber meets the road with new ORNL carbon, battery technologies
Breaking benzene
Borek
#2
Mar26-12, 11:02 AM
Admin
Borek's Avatar
P: 23,574
Beer-Lambert is a way to go.

Quote Quote by FaNgS View Post
Is it correct to take 100μL of my sample and diluting it with 900μL of Methanol then multiplying by a factor of 9 to obtain the actual concentration?
I would prefer to dilute in a separate volumetric flask. Not only you will get much better accuracy of the final volume, but you will be also able to thoroughly mix the sample with the solvent, which you won't be able to do in a cuvette.

Make sure your cuvettes survive methanol bath.
Ygggdrasil
#3
Mar26-12, 02:03 PM
Other Sci
Sci Advisor
P: 1,389
Quote Quote by FaNgS View Post
Is it correct to take 100μL of my sample and diluting it with 900μL of Methanol then multiplying by a factor of 9 to obtain the actual concentration?
You would have to multiply by a factor of 10 to get the correct concentration. For example, if you dilute 100 ÁL of a 1 M sample in 1 mL total solution, you would have a 0.10 M solution, not a 0.11 M solution.

Also, if you are measuring absorbance in the UV range, you would need to use quartz cuvettes as plastic cuvettes absorb UV light (not to mention the possibility of the MeOH leaching UV active compounds from the plastic).

Absorabance values are generally reliably in the range from 0.1-1.0. If your samples are too dilute to get absorbance values in this range you may not be able to get a reliable measurement of the concentration.

FaNgS
#4
Mar27-12, 04:16 AM
P: 92
Using UV-Vis Spectroscopy to determine concentration

I asked a colleague who has been using the UV-Vis in the lab and he told me that I should prepare different standard solutions of varying concentrations.

My initial adsorption experiment solutions are at 200ppm and 400ppm so according to him I should first prepare standards at 50ppm, 100ppm, 150ppm, 200ppm, 250ppm, 300ppm, 350ppm, 400ppm.
Then for each standard I will measure the absorbance using the machine and create a straight line plot of Absorbance (y-axis) vs. Concentration (x-axis). From the plot I will get the equation of the line in the form y=mx+c where y is the absorbance, m is the slope, x is the concentration and c is the y-intercept.

Then I will measure the absorbance of my samples, and using the equation of my plot I will calculate my sample concentrations. According to him this is one form of the Beer-Lambert law. Is this a proper procedure to follow?

Another thing he mentioned is that I should specify the wavelength at which the machine will detect the peaks based on the wavelength of the compounds I am interested in, which are toluene and xylene. So I've been searching and I was able to get some UV wavelengths reported in the Combined Chemical Dictionary Database but there are several wavelengths reported for each compound.

For instance this is what is reported for Toluene but which one should I use?:

[neutral]λmax189 (log ε4.75) ;206 (log ε3.91) ;248 (log ε2.1) ;254 (log ε2.2) ;256 (log ε2.3) ;258 (logε2.3) ;260 (log ε2.3) ;262 (log ε2.4) ;265 (log ε2.2) ;268 (log ε2.4) (heptane)
Borek
#5
Mar27-12, 05:15 AM
Admin
Borek's Avatar
P: 23,574
Quote Quote by FaNgS View Post
I asked a colleague who has been using the UV-Vis in the lab and he told me that I should prepare different standard solutions of varying concentrations.

My initial adsorption experiment solutions are at 200ppm and 400ppm so according to him I should first prepare standards at 50ppm, 100ppm, 150ppm, 200ppm, 250ppm, 300ppm, 350ppm, 400ppm.
It is not incorrect, but I would prefer to finish at - say - 600 ppm. This way you will cover not only concentrations below those you expect, but also those slightly higher than expected. Better safe than sorry.

Then for each standard I will measure the absorbance using the machine and create a straight line plot of Absorbance (y-axis) vs. Concentration (x-axis). From the plot I will get the equation of the line in the form y=mx+c where y is the absorbance, m is the slope, x is the concentration and c is the y-intercept.

Then I will measure the absorbance of my samples, and using the equation of my plot I will calculate my sample concentrations. According to him this is one form of the Beer-Lambert law. Is this a proper procedure to follow?
Looks OK to me, yes, this is an application of Beer-Lambert law.

Another thing he mentioned is that I should specify the wavelength at which the machine will detect the peaks based on the wavelength of the compounds I am interested in, which are toluene and xylene. So I've been searching and I was able to get some UV wavelengths reported in the Combined Chemical Dictionary Database but there are several wavelengths reported for each compound.

For instance this is what is reported for Toluene but which one should I use?:

[neutral]λmax189 (log ε4.75) ;206 (log ε3.91) ;248 (log ε2.1) ;254 (log ε2.2) ;256 (log ε2.3) ;258 (logε2.3) ;260 (log ε2.3) ;262 (log ε2.4) ;265 (log ε2.2) ;268 (log ε2.4) (heptane)
It depends. Usually we select the wavelength that grants the highest sensitivity, but sometimes it is better to select different wavelength, depending on the impurities present.
FaNgS
#6
Mar27-12, 06:22 AM
P: 92
Quote Quote by Borek View Post
It depends. Usually we select the wavelength that grants the highest sensitivity, but sometimes it is better to select different wavelength, depending on the impurities present.

What defines the sensitivity?
The chemicals are high grade purity, with methanol at 99.99%, toluene and xylene at 99.44%.
chemisttree
#7
Mar27-12, 11:44 AM
Sci Advisor
HW Helper
PF Gold
chemisttree's Avatar
P: 3,724
Lambda max is where you will have best sensitivity. Should be about 254-270nm.
The molar extinction coefficient for toluene has been measured as 2,864 cm-1/M at 261.8 nm.
FaNgS
#8
Apr1-12, 12:09 PM
P: 92
This is a follow-up question on my progress. I was able to obtain 2 quartz cuvettes (10.00mm) and so far the absorbance measurements giving me consistent results with my standard solutions.

However, once I reach my maximum concentration solution of 450ppm the absorbance given my the UV-Vis machine is just over 1.0. Please correct me if I am wrong, but doesn't an absorbance of 1 mean that my solution is completely black and no light can pass through? (If this is true then my results are all rubbish!!! :S )

I found this site which explains UV-Vis and the Beer-Lambert law: http://www.chemguide.co.uk/analysis/...erlambert.html
So according to what is written, an absorbance greater than 1 is possible but only in "extreme cases". So what exactly is happening? Because my solutions are colorless (methanol solvent with ppm levels [50ppm-450ppm in 50ppm increments] of individual benzene, toluene or xylene).

Please note that I am still only working with standard solutions and I have not YET used my adsorption experiment samples (which would require dilution to cover enough of the cuvette for the light to pass through the solution). So are my results rubbish?
Ygggdrasil
#9
Apr1-12, 12:39 PM
Other Sci
Sci Advisor
P: 1,389
Absorbance is defined as A = -log(T), where T is the fraction of light transmitted by the signal. So an absorbance of 1.0 corresponds to a 10% of the light being transmitted by the signal (i.e. 90% of the light is absorbed).

Note that because you are likely looking at absorbance in the UV, it is perfectly reasonable for a high absorbance sample to appear colorless as our eyes cannot see UV light.
FaNgS
#10
Apr30-12, 05:43 AM
P: 92
I am back with my results from my adsorption experiment but they don't make any sense and are not logical (to me at least)! I have uploaded my results as a pdf file on google docs, I hope someone can have a look https://docs.google.com/open?id=0B8v...zd3UE5QY2dKMFk

Let me explain my experimental procedure: I initially prepared different solutions of known concentrations of Benzene in Methanol from 50ppm up to 650ppm in 50ppm increments. Using the UV-Vis spectroscopy I measured the absorbance of each solution.

I then plot a curve of absorbance vs concentration to represent my calibration/standard curve. I then used a spreadsheet to generate a linear fitting trendline and an equation fitting the trendline to the data points. This procedure was repeated for Toluene in Methanol solutions (50ppm -450ppm) and Xylene in Methanol solutions (50ppm - 450ppm).

For the adsorption experiment, in 3 separate glass bottle I poured out 40mL of 200ppm Benzene solution (the same one I used earlier for my calibration curve) and in each bottle I added a different activated carbons. I have 3 different carbons AC40, C38 and C40. Similarly, this done for 400ppm solution of Benzene.This was all done, in the same manner for Toluene and Xylene.
(So I would have 3 bottles with 200ppm benzene solution each with a different carbon, 3 bottles with 400ppm benzene each with a different carbon. 3 bottles of 200ppm Toluene solution each with a different carbon and so on..)(In total I had 18 bottles).

At 12 hour increments (from 12 to 36hrs) I measured out 1mL from each bottle and collected it in a small vial which was later measured in the UV-Vis spectroscopy to determine the concentration change with time.

So if you have a look at the pdf, each page has a table of data and a plot of absorbance vs concentration. The plot represents the calibration curve I did using known concentrations of solutions I prepared. Using the equation generated from the calibration curve and the absorbance measured I calculated the concentration.

But in some cases the concentration calculated is much higher than my initial concentration and in other cases the concentration would increase significantly after a 12-hour period.

Any ideas? I am currently completely lost and don't know in what direction to proceed.
qalomel
#11
May1-12, 04:24 PM
P: 11
That is strange. Either there are matrix influences in your sample or it's just random error. Do all the peaks occur at the same wavelengths for each sample between the different time frames?
FaNgS
#12
May12-12, 01:35 AM
P: 92
Each chemical (Benzene, Toluene, Xylene) has a lambda max which gives the maximum absorbance. I measured the absorbance at these wavelengths to get the highest absorbance.

Edit: So for Benzene literature reports a lambda max of 255nm, 262nm for Toluene and 265nm for Xylene. I also confirmed that at these wavelengths the UV Spectroscopy I was using did actually measure the max absorbance possible.

I'm really sure what is influencing the measurements. I re-measured my standard solutions and although the absorbance was slightly different (I believe due to vapor loss since all the chemicals involved are organic) which also caused a slight change in my calibration curve equation, when I re-calculated my concentrations it did not make a big difference and the values were still much higher than the starting concentration or would be greater than the previous time-frame.

A small side track here but what methods are usually used in order to determine concentrations of unknown organic solutions? What equipment can be used?
Ygggdrasil
#13
May12-12, 10:27 AM
Other Sci
Sci Advisor
P: 1,389
A few notes:

1) Your experimental measurements should really be done with multiple independent samples (e.g. in triplicate) so that you can get some measure of the variability of your measurements. Maybe the increase you see is not a real increase, but just reflects the standard deviation of your measurement.

2) Maybe it would be worth including a step to ensure that none of the activated carbon is present in the solution you measure on the spetrophotometer (e.g. centrifuge the samples prior to putting them in the cuvette). Particulates in your solution can increase the absorbance of the solution as they scatter light.

3) The standard method to determine the concentrations of unknown organic solution would probably be gas chromatography.
FaNgS
#14
May12-12, 10:55 AM
P: 92
1) i.e having replicates of each batch/experimental solution?

2) I did filter my solutions before adding them to the cuvette

3) Are there any other non-standard/conventional methods? What about GC-MS?
Ygggdrasil
#15
May12-12, 11:01 PM
Other Sci
Sci Advisor
P: 1,389
1) Yes, perhaps if possible, take multiple aliquots from your standard solution, incubate each with your activated carbon and take measurements from each.

2) How are you doing the filtration? I could imagine vacuum filtration could cause problems due to the volatility of some of the compounds.

3) GC-MS would work just as well a GC alone.
MrSid
#16
May16-12, 10:11 AM
P: 81
I was wondering if you are doing background blank extractions of your matrix. Also the purity of your MeOH and other solvents should be confirmed for UV spectrophotometric use, if you are to see these small quantities of aromatics. Obviously the UV bands for toluene, and benzene won't be able to be resolved so your data will only be for total aromatics.

You will have to use syringe filters that won't have extractable impurities or adsorb the analytes.

In principle, GC-MS is the way these studies are done and the activated carbons are washed and extracted for blanks before and after the studies. The method of Standards Additions is used to account for matrix effects, and an external standard (like Tol D8) is used to differentiate the numbers- triplicates are are also useful, and the standards have to bracket the unknowns.


Register to reply

Related Discussions
Determine the concentration/purity of Caffeine in a precipitate Chemistry 3
Why do ions flow from high concentration to low concentration? General Physics 2
[Chemistry] Determine the concentration of h2o2 by redox titration! Biology, Chemistry & Other Homework 5
Determine the impurity doping concentration Engineering, Comp Sci, & Technology Homework 0
Titration method to determine HCl concentration in a mixture Chemistry 7