HPLC 101 - Help me make a calibration curve

[Deppizzymo]

Ok... So we got our Agilent 1100 up and running on a Waters NH2 column producing good peaks for general sugars analysis. I am wanting to make a calibration curve for 6 components (glucose/galactose/fructose/lactose/sucrose/maltose) at varying levels. Our client is a waste stream recycling group that combines lots of food waste / byproducts into animal feed. I get both raw materials and mixed finished products and they are all over the map for sugar content. I am going to start by trying to come up with a good curve for their finished products... these range from 0-5% galactose, 5-10% fructose/glucose/maltose, 20-30% lactose and 50-60% sucrose. I have never done this stuff before but I created a calibration curve by making my own mix of pure sugars from a kit as follows...

0.6g sucrose, 0.1g maltose, 0.3g lactose, 0.1g fructose, 0.1g glucose, 0.05g galactose and diluting it to 50mL to produce a 2.5% stock solution. I then diluted this stock solution and made 5 dilutions: 1 parts stock:5 parts water, 2:5, 3:5, 4:5, and the full strength solution. I did three injections of each point, 1:5x3, 2:5x3, 3:5x3, 4:5x3, 5:5x3. I created my curve by using % amount and assigned the values as follows...

full strength stock
60% sucrose, 30% lactose, 10% maltose, 10% fructose, 10% glucose and 5% galactose

4:5 solution
48% sucrose, 24% lactose, 8% maltose, 8% fructose, 8% glucose and 4% galactose

3:5 solution
36% sucrose, 18% lactose, 6% maltose, 6% fructose, 6% glucose and 3% galactose

2:5 solution
24% sucrose, 12% lactose, 4% maltose, 4% fructose, 4% glucose and 2% galactose

1:5 solution
12% sucrose, 6% lactose, 2% maltose, 2% fructose, 2% glucose and 1% galactose

Doing this I had great correlations of 0.99 or better on every sugar curve.

I am here to ask... does this all make sense/seem logical/how you would do it? I am here because after analyzing our sugar QA sample (prepped 1.25g->50mL) that we test weekly on our current sugar method of spectrophotometry/enzyme kits, I am getting pretty significantly different results than we expected. They are all elevated... results are

3% glucose/36% sucrose/2% maltose/21% lactose 64% total sugar on the HPLC
vs
1% glucose/28% sucrose/1% maltose/15% lactose 45% total sugar with the kit

I am certain it has to have something to do with my diluting/telling the system what the sugar %amounts are but I don't see where I might be wrong. "you don't know what you don't know" sort of situation.
I really appreciate any help I can get!

[TylerSmith123]

Hi Depizzymo,

Let me ask you this preliminary question, how are you calculating these percents?
I assumed you were using a weight/volume measurement, but I am having trouble getting to your results.

As you state, you use, for example, 0.6g of sucrose in 50 ml of diluent,
0.6 g / 50 ml x 100 = 1.2%
Commonly, w/v is reported in g/100ml so I will simply do the calculation for that:
0.6g / 100 ml x 100 = 0.6%.
However, assuming we carry on with the former calculation then: suc: 1.2%, malt: 0.2%, lact: 0.6%, fruct: 0.2%, and galact: 0.1% (all in 50 ml), do not add up to the 2.5% stock solution that you indicate, but a 2.3% stock solution.

In all honesty, I think I've been staring at some of these percent values too long and my brain is starting to numb haha. Hopefully this is the simple solution that you are looking for, if not it probably is your range. I see that your test sample (1.25 g) is outside of your calibration area, I would make a stock solution that is, at maximum, comfortably away from the maximum that you are testing in actuality. This will mean that the values that you have for your calibration curve actually fit the region of your field samples and you can pass/fail based on those values.

I hope someone else can hop in and give you a more concrete answer, but hopefully I helped a little!

[Deppizzymo]

Hi Depizzymo,

Let me ask you this preliminary question, how are you calculating these percents?
I assumed you were using a weight/volume measurement, but I am having trouble getting to your results.

As you state, you use, for example, 0.6g of sucrose in 50 ml of diluent,
0.6 g / 50 ml x 100 = 1.2%
Commonly, w/v is reported in g/100ml so I will simply do the calculation for that:
0.6g / 100 ml x 100 = 0.6%.
However, assuming we carry on with the former calculation then: suc: 1.2%, malt: 0.2%, lact: 0.6%, fruct: 0.2%, and galact: 0.1% (all in 50 ml), do not add up to the 2.5% stock solution that you indicate, but a 2.3% stock solution.

In all honesty, I think I've been staring at some of these percent values too long and my brain is starting to numb haha. Hopefully this is the simple solution that you are looking for, if not it probably is your range. I see that your test sample (1.25 g) is outside of your calibration area, I would make a stock solution that is, at maximum, comfortably away from the maximum that you are testing in actuality. This will mean that the values that you have for your calibration curve actually fit the region of your field samples and you can pass/fail based on those values.

I hope someone else can hop in and give you a more concrete answer, but hopefully I helped a little!

So sorry... I totally forgot to include the 0.1g glucose. I am trying to tell the instrument that if the peak is "x area", report this as 30% glucose. My samples from our customer are dry powders that are say, 30% glucose as-is. I want to dilute this so I can run it at an appropriate range on the instrument and have it report out 30% glucose. Does this make sense?

[H.Thomas]

A few thoughts.

Your percentage calculations are wrong. Your calibration sample weighs 1.25g and contains 0.6g sucrose. That makes 48%, not 60.

Of course you can calibrate the method in percents, if you always weigh out exactly 1.25g diluted in exactly 50 ml water. I would not recommend it though.

You probably have some kind of spreadsheet. Calculate the concentrations:

Sucrose 0.6(how many zeroes?) g in 50 mL =12.0 g/L

Your calibration dilutions contain 2.4 g/L 4.8 g/L 7.2 g/L 9.6 g/L sucrose

Calibrate with these concentrations. If you now measure a sample, you take the weight (e.g. 1.256 g in 50 mL) and the measured concentration (say 10.1 g/L sucrose) and calculate:

10.1 g/L * 50 mL / (1000 mL/L * 1.256 g) = 0.40 or 40%.

If your sample happens to contain more than 50% sucrose, i.e. you're out of the calibrated range, you can weigh less or dilute the solution.

That's the way I would do it.