Cooling A Flask?

So, you have a flask (maybe it contains some chemical) and you want to cool the flask. The obvious answer is to place it in a refrigerator. However, you want to keep it on a magnetic stir plate, which every well stocked kitchen has on hand.  You do have a few options, the first is to place the flask and the stir plate into the refrigerator, but then the power cables might block the door… Or, you can use a cooling water bath, like a kitchen ice cream maker, with ice in a container around the thing to be cooled (messy).  Unless you have a temperature controlled recirculating water bath, you won’t be able to control the temperature.

There is another option.  A Thermoelectric or Peltier Cooling Chip AND our previous temperature controller project.  These work by applying electricity to a series of electrical junctions. When the current flows it moves the heat with it – discovered back in 1834. It is moving entropy from one area to another, this change is reflected in the measured temperature – items at lower temperature tend to have less disorder.  So, we can use one of these cooling chips for our flask.  Now, how big of a chip to buy?

We can apply a bit of heat transfer.  Let us assume the room is a nice warm 23°C (or 74°F). There is not wind in the kitchen, so we can say the heat transfer coefficient from the flask wall to the room air is about 25 watts per meter squared per degree centigrade. This number indicates the amount of heat (25 watts) that will exchange to/from the air/flask for each unit of exposed area and degree of temperature difference.  So, the bigger the difference in temperature the more heat will want to move between the two items, and more area exposed the more active exchange points exist.

Since we are cooling the flask (lets pick 10°C or 50°F (a refreshing temperature) we can find the amount of heat (Qin) that will be coming into the flask from the warm air.  This is simple, make your units work. Qin equals the difference in temperature (23°C – 10°C) times the flask exposed area in squared meters (0.04 m2). Qin is about 13 watts.  So, we a chip that can move at least 13 watts.

Now, if we are using the chip to cool the flask the heat needs to go somewhere…. So, we can put the chip under the flask between the stir plate top and the metal/ceramic plate (use a bit of paper towel to form an insulator).  Question? is the plate the big enough to handle the heat from the flask with out getting too hot?  Let us assume the hottest we want the plate top to get is 50°C (hot bath water).  We can do the same calculations, Q to room equals the area of the stir plate exposed to the room air (0.04 sq meters minus the space covered by the flask 0.01 sq meters) times the coefficient of transfer (25 water per sq meter per degree C) times the driving force difference in temperature (50°C – 23°C).  This is about 24 watts… We can apply 24 watts to the top of the stir plate before it gets too hot.  Good!  We are only moving 13 watts from the flask.  If this number was too low, the project wouldn’t work or the stir plate would get too hot to be safe. If you have that that problem, you might need cooling fins, a fan, and a bigger stir plate.

Now we know that we can buy a cooling chip that can move at least 13 watts and keep our flask nice and cool.  Instructions about building the actual temperature controller are over here in another post.

Coolling flask calculations




Publication – Lowering the cost of cellulase

There is a new publication that was just accepted for print regarding the production of cellulase. This article is entitled “Nutrient control for stationary phase cellulase production in Trichoderma reesei Rut C-30″ and will appear in the journal Enzyme and Microbial Technology.

Cellulase is normally produced by a fungus, Trichoderma ressei.  Most of the time the cells continually grow and make enzymes needed to consume cellulose (what trees are made of – think mold growing on a dead log).  In this study, we show that by using phosphorus limitation, we can keep the fungal cells from growing and yet remain healthy.  This condition allows them to produce cellulase enzyme without growing in size.  For industrial cellulase production with less growth, there are less cells that need to be sent to a waste stream and less nutrient is used to make wasted biomass.  These improvements allow us to cut the costs of cellulase production by 60% and make the process more efficient.

The full article and results will appear in Enzyme and Microbial Technology, you can view the pre-print edition here.

Calibration Data

A question that I seem to receive frequently: “Can you help me with the calibration?”

So, you have a sample.  And, there is some chemical in there (like glucose) and you want to know how much glucose is in the sample… maybe you have some idea, it should be less than something, right?  Were do you start?

Step 1: You select a test method.  You might choose HPLC, UV-VIS, FTIR, etc.
Step 2: Calibrate your machine.
Step 3: Measure your sample
Step 4: Ask, does the result make sense?

Calibrations are done to correct for machine errors and produce correlations that relate things (usually a signal value to a concentration).  Most lab equipment in common labs work with Light.  A UV-VIs (Spectrophotometer), Refractive Index detector, FTIR, Florescent detector, etc, all use light. The light passes through (or sometimes reflects off) your sample. The amount of the light that does or doesn’t make it past your sample is related to the concentration.

Most detectors have a “Linear Response Rage.”  This is the operating condition where the signal is directly related to the amount of material it is measuring.  And they have a “Lower Detection Limit” or the smallest amount that they can detect.  There is also the “Sensitivity” that is how well it can determine the small differences.

If your sample blocks all the light, the detector maxes out the signal.  No matter how much more you put in the way of the light, if all the light is blocked – all the light is blocked – your signal stays the same… at the max. You will need to dilute your sample and reduce the number of molecules that are blocking the light.

Now, if your sample is really dilute or the light is really bright that it doesn’t even notice the sample molecules then the detector says you have nothing there… This is the “Lower Detectable Limit.” The lowest concentration of molecules that are needed for the machine to notice. You might need to concentrate your sample, that is dry off the solvent to make is more concentrated.

Most machines are tuned towards the low end. That means they will detect a very low concentration of molecules. That setup sacrifices the detector’s upper limit. It will be saturated (all the light is blocked) at lower amounts of sample. You can fix this issue by diluting your sample to make it fit within the range that the machine can “see” well.  It is much easier to dilute a sample then try to dry it or concentrate it.

We call the typical operating range the “Linear Range.” If you plot the signal as a function of sample concentration, you will actually see a drop off in the signal on the far left, a zone that looks like a line, and a plateau where the light is blocked. See the picture below, if the signal is above 0.65 or below 0.1 the data can’t be used. Also, keep in mind that every machine, test dye, and procedure will require its own calibration curve… don’t use this one here for your data.

Operating ranges for a detector. You want to work in the linear range. That area is the most accurate.

Operating ranges for a detector. You want to work in the linear range. That area is the most accurate.


So, how do you make the calibration standards?

Maybe you are measuring something easy, glucose… a standard HPLC + Refractive Index Detector (RID) and an H+ Column will work for this test.  Your samples are all in water.

Step A: Make a stock solution that you are absolutely certain of the concentration (that is grams per liter) and not amount (grams). These tests work on concentration not absolute amount. To make this solution, use a volumetric flask (these have a line etched in the glass to ensure the correct volume when full). Measure out an amount that your balance can detect (something like 1 gram) and put it into the flask. Now add water (at the temperature stated on the flask – the flask should state the volume of water at the line +/- the accuracy and the temperature for that volume). Now, add the water to the line. Cap. Mix. If you used a 100 mL flask and 1 gram of glucose you have a 10 g/L glucose standard. — By the way, bacteria and fungi eat sugar so, this standard won’t last more than a day or two!

Step B: Dilute the stock solution into known concentrations.  The simplest way is by factors of 2.  That is you take 10 mL of 10 g/L standard and add 10 mL of water.  Now you have a 5 g/L standard.  Take 10 mL of that into a new container and add another 10 mL of water. Now you also have a 2.5 g/L standard. Do this a few times… to create a set 10, 5, 2.5, 1.25, 0.625, 0.313, 0.156  g/L  You need to use accurate volume measuring devices to do the dilutions. DO NOT WEIGH THE LIQUIDS (you added sugar to the stock, that changes the density and weighing it isn’t accurate unless you measure the density and apply a correction – Likewise make sure all the liquids are at the same temperature when you measure out the volume).

Step C: Test these samples and your unknowns (AT THE SAME TIME). If you are using a colormetric, or wet chemistry where your are measuring UV-VIS absorbance, record the absorbance number. For HPLC, you record the area under the peak (not the peak height). Since, test dyes, columns, light bulbs, your mood, can change with time, you want to test your standards at the same time you test your unknowns… no cheating…

Step D: Plot your data. You plot should look like the one above. If you use Excel, Gnumeric, OpenOffice, Apple Numbers, add a trend line to the linear range. The equation is called your calibration curve (even though it is linear – is called a curve). Normal convention puts the concentration as the x-variable (you set the concentration first) and the signal (or area) as the y-variable (that is dependent on the sample concentration). In practice, it is easier to have the signal (or area) as the x-variable and the concentration as the y-variable (this saves you from doing the algebra).

Step E: Use the calibration curve to convert the area or absorbance of your unknown to the concentration.  DONE!


I have included a Excel Calibration Workbook to help you.

Keep in mind that all these methods are based in concentration, not absolute amount. If you need to get to an absolute amount, you need to multiply by a volume term. So, you have a bottle of sugar that is 200 mL… (that is 0.2 L) and you used 10 uL to inject for the HPLC test.  The HPLC says it is 2.5 g/L — so, 2.5 g/L * 0.2 L = 0.5 grams sugar in the bottle + some amount of water.

Significant figures & Number of Digits: These are limited to the lowest accuracy device you used. So, if your volumetric flask was only good to 2 digits (+/- 0.02 mL at 20ºC ) your results are only good to 2 places 0.01 (you can’t go about adding more digits). If you think you need more accuracy you need to buy more accurate measuring devices… and then do multiple tests to find out how reproducible you or your machine is… +/- 0.001 is just about the limit for most normal things… if you see extra digits in someone’s data… question them as to why?

If you think you need help, message me on Research Gate.

Good Luck!