Malt Titration (finding pHDI, a, b and c)

Homebrew Talk - Beer, Wine, Mead, & Cider Brewing Discussion Forum

Help Support Homebrew Talk - Beer, Wine, Mead, & Cider Brewing Discussion Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

ajdelange

Supporting Member
HBT Supporter
Joined
Aug 5, 2010
Messages
11,959
Reaction score
2,747
Location
McLean/Ogden
This is in response to a request for tips on how to titrate a malt sample and determine a, b and c from the titration data. The request was that this be a Sticky. I'm not sure that is justified as I doubt many will want to undertake malt titration.

The procedure we are going to try to simplify for home brewing here is found in my paper "Predicting and Controlling Mash pH Using Simple Models for Mash Component Acid/Base Characteristics" MBAA TQ vol. 52, no. 1 2015 pp 3 - 12.

The simple model for malt is

∆Q(pHz) = a*(pHz - pHDI) + b*(pHz - pHDI)^2 + c*(pHz - pHDI)^3

∆Q(pHz) is the number of mEq of protons that must be added to a kg of malt mashed in a reasonable amount of DI water to bring the pH of the mash to pHz. If ∆Q(pHz) > 0 then one adds the protons by adding acid. If ∆Q(pHz) < 0 then protons must be removed, i.e. absorbed, by adding base.

pHDI is the distilled water mash pH for this malt. a, b, and c (called a1, a2 and a3 in the paper) are coefficients which describe, along with pHDI, the malts acid/base characteristics. To determine them we add acid and base in different amounts to small portions of the malt in question mashed with DI water and measure the pH. The first pH measurement is made with no acid or base addition and the resulting pH is pHDI. We then plot the acid or base additions vs the measured pH and find values of a, b, and c which produce a curve which best fits the data.

Some example data is shown in the plot below where each circle represents a measurement. It's based on the Weyermanns pilsner malt curve in the paper but has been played around with to illustrate a few points which we'll get to. For starters, you measurements won't come out on nice whole tenths of a pH. The job is to find the curve which best fits the data in the rmse sense. We'll have Excel do that for us but first let's find out how to get the measurements.
TitrEx.jpg

What You Will Need

1)0.1 N Sulfuric Acid (Hach)
2)0.1 N Sodium Hydroxide (Hach)

We need to measure out protons and proton absorption capacity in mEq. An 0.1 N solution of sulfuric acid supplies 0.1 mEq/mL an an 0.1 N solution of sodium hydroxide absorbs 0.1 mEq/mL. Obviously you can get 0.1 N solutions from many places or make them up yourself if you know how and know how to standardize them. Hach sells them in 1 liter bottles and all the work is done for you but you know what the catch is - you have to pay for that.

3)Some way to measure small quantities of these solutions accurately. Syringes that can be read to tenths of a cc should do. A buret is better, an adjustable pipetter better still and a Hach Digital Titrator best (IMO)*.

4)A means to finely grind small quantities of malt. I use a Waring spice grinder.

5)A means to weigh out 50 grams of malt accurately

6)A source of DI water

7)A means to measure out 100 mL of water (great accuracy is not required here)

8)Beakers (ideally stainless steel) to hold the mini mashes (50 grams of powdered malt with 100 mL of water)

9)A good pH meter calibrated frequently

10)A means for holding the beakers at about 50° C.

11)pH 4 and 7 buffers,

Procedure:

1) Transfer small quantities of pH 4 and 7 buffer into containers which can be immersed in the water bath and place them there.
2) Measure out 100 mL DI water and place the container in the water bath at 50 °C
3) Finely grind somewhat over 50 grams of malt. Mix the grind thoroughly so the husks are evenly distributed throughout the flour. Measure 50 grams as accurately as possible and transfer to a beaker.
4) Place the beaker in the water bath (a weight will probably be necessary to keep it submerged)
5) Wait until the temperature of the DI water is 50 °C (or close to it)
6) Mix the water with the malt stirring thoroughly. Continue to stir periodically until 25 minutes have passed.
7) During the waiting period calibrate the pH meter using the 50 °C buffers.
8) Measure the pH (yes, at 50 °C). Be sure the reading is stable before accepting it (at least a couple of minutes)
9) Remove the beaker from the water bath and cool it to room temperature.
10) When the mash is at room temperature measure the pH again.

11) Look at the pHDI reading obtained at 50 °C. in the example data set it was 5.49. This is right in the middle of the mash pH range so you will be using the parameters to calculate ∆Q's at pH's above and below this value. Thus you will have to titrate with both acids and bases. It's good to have as many data points as possible, of course, but this is pretty tedious work so you might decide to shoot for measurements separated by 0.2 pH. You know that the buffering of most malts is about -40 mEq/kg•pH so that to raise the pH of a mash of this malt to 5.69 you would need to absorb something like 40*0.2 = 8 mEq/kg. You are mashing 50 gram (equals 1/20 kg) samples and so would require 8/20 = 0.4 mEq proton absorbing capacity. You have 0.1 N NaOH solution which absorbs 0.1 mEq/mL and thus would need 4 mL of that. Accurately measure 4 mL of 0.1 N NaOH into about 90 mL of DI water
12) Repeat from Steps 2 - 8 except instead of measuring out 100 mL of DI water in Step 2 make up the water NaOH mix from Step 11 to 100 mL.

From here on you continue taking data, plotting each point as it comes in, until your data set looks something like the set of circles in the picture above. To collect lower pH points use exactly the same reasoning as given in Step 11 but for acid instead of base. IOW add 4 mL of the 0.1 N H2SO4 to 90 mL of water and use that.

Obviously, as noted above, this is going to get tedious pretty quickly. About the only recommendation I have there is to see if you can handle more than one titration at a time. That way you get three points in an hour instead of only 1 but you have to figure out a protocol that lets you juggles three mashes at once. Preparing 3 malt samples and 3 water/acid/base mixes and then mixing water and malt staggered 3 - 4 minutes apart seems to work for me.

Note that your measured pH's will not be on nice even tenths of pH's nor will they be evenly spaced in 0.2 pH increments above and below pHDI. This is because, of course, the malt you are measuring won't have a = -40 mEq/kg (what a actually is is what we are trying to find out.

I'm going to stop here for fear of overflowing the size limitation. To be continued.

*This is in part because the digital titrator is accurate, easy to use, can be used for other things such as measuring alkalinity and hardness and because the NaOH titrant comes in a cartridge which is sealed in a plastic bag thus protecting it from air (CO2)
 
Last edited by a moderator:
We now turn to the subject of processing the mEq/kg vs pH data to get estimates of a, b and c. In words this is a matter of using a curve fitting routine which calculates the values of a, b and c which minimizes the sum of (∆Q - mEq/kg)^2 in which ∆Q is calculated from a, b, c, pH and pHDI for each measured point. Curve fitting routines abound which will fit a + b*x + c*x^2... but not so many that will fit a*(x - x0) + b(x-x0)^2... I use IGOR which made the plot in No. 1. There is probably an Excel function for at least the polynomial fit but I'm going to present a slightly different approach which, I hope will lend some insight.

Open a fresh Excel workbook. Copy and Paste the following codes into Columns A through D:

Column A:
pHdi
a
b
c
pH's
4.8
5.485
5
5.2
5.4
5.6
5.8

Column B:
5.485
-35.5389729490434
14.4086232829913
-12.7481925660599
mEq/kg
35
0
23
10
4
-5
-10

Column C:




∆Q Calculated from a, b, c
=$B$2*(A6-$B$1) +$B$3*(A6-$B$1)^2 +$B$4*(A6-$B$1)^3
=$B$2*(A7-$B$1) +$B$3*(A7-$B$1)^2 +$B$4*(A7-$B$1)^3
=$B$2*(A8-$B$1) +$B$3*(A8-$B$1)^2 +$B$4*(A8-$B$1)^3
=$B$2*(A9-$B$1) +$B$3*(A9-$B$1)^2 +$B$4*(A9-$B$1)^3
=$B$2*(A10-$B$1) +$B$3*(A10-$B$1)^2 +$B$4*(A10-$B$1)^3
=$B$2*(A11-$B$1) +$B$3*(A11-$B$1)^2 +$B$4*(A11-$B$1)^3
=$B$2*(A12-$B$1) +$B$3*(A12-$B$1)^2 +$B$4*(A12-$B$1)^3

Column D:




(∆Q - mEq/kg)^2
=(B6-C6)^2
=(B7-C7)^2
=(B8-C8)^2
=(B9-C9)^2
=(B10-C10)^2
=(B11-C11)^2
=(B12-C12)^2

=SUM(D6:12)

When finished you should have a spreadsheet that looks like this:
Untitled21.jpeg


Enter your measured pH DI into B2 an your other measured pH data into columns A and B. Note that a base addition is entered as a negative number (the malt has a negative proton deficit, i.e. a proton surfeit at this pH) whereas acid additions are entered as positive proton deficits (this much acid is required to reach this pH). The numbers in col. B for a, b and c can really be anything but putting -40 in B2 may speed convergence of the Solver. The number in column C are values of ∆Q calculated for the corresponding pH in col. A, pHDI in B1 and the a, b and c values in Col. B. The values in Col. D are the squares of the differences between the ∆Q values and he amounts of acid or base we actually added. Cell D14 is the sum of those squares. We wish to find a, b and c that minimize that sum. We have the Excel Solver do that. Go to the Tools menu and invoke the Solver configured thus.


Untitled22.jpeg


Be sure the Make Unconstrained Variables Non-Negative box is NOT ticked. Upon pressing Solve the Solver will grind away, fiddling with B2, B3 and B4 until it gets the smallest value of D14 it can. Note that this is going to be 0. Examination of the curve in No. 1 or the Column D numbers shows that there is no 3rd order Taylor series which captures each of our measured points perfectly. This is true in the real world but in preparing data for this example I intentionally introduced extra error into the point at pH 5.2. This is to illustrate that a single "bad" data points effect can be somewhat nullified if more data points are taken and is an argument for as many data points as you can stand. The minimum required is the pHDI measurement and two others. If there are only 3 points the curve will go right through all three points. The effect of this can be seen on the plot in No 1. Picture the curve going through that 5.2 pH point.

Those interested are bound to have questions. Shoot.
 

Attachments

  • Untitled21.jpeg
    Untitled21.jpeg
    47.3 KB · Views: 64
Last edited:
An important aspect of this I forgot to mention in Nos. 1 & 2 is that the 0.1 N NaOH solution won't stay at 0.1 N because it will pick up CO2 from the air which will react with the NaOH:

CO2 + H2O + 2Na+ + 2OH- <--> 2Na+ + CO3-- + 2H2O

CO3-- isn't as strong a base as 2 OH- so you will need a bit more of the solution to get the same amount of proton absorption. To check on this carefully measure out 4 mL of 0.1 N H2SO4 and add 4 mL of the NaOH solution. Mix well. The pH of the resulting solution should be between 6 and 8. If it is less than 6 add tiny amounts of additional NaOH until 7 < pH < 6. Record this extra amount and compute (4 + extra)/4. The normality of your solution is now 0.1*(4/(4 + extra)) and you will need to adjust all base amounts calculated base on 0.1 N upward by the factor (4 + extra)/4. Don't continue to use this NaOH solution if (4 + extra)/4 > 1.05,
 
AJ some months back you provided me with the instructions below when asked if I could use Lactic Acid for the titration testing.

"If you add 1 cc of 88% lactic acid to 100 mL of DI or RO water you will have a solution that is pretty close to 0.1 N so you would use 1.6 mL of that in order to get 0.16 mEq."


I plan to mix up the Lactic Acid based 0.1N solution as you described and use it in my testing.
 
The .1 N stuff I'm recommending is pretty benign. You shouldn't have to pay hazardous material shipping fees with it. Should you accidentally drink a liter of the acid you could relieve the acid stomach with about 1.5 tsp bicarbonate of soda. Be prepared, however, to belch up about 2 L CO2.
 
Is there any way to standardize unknown strength cleaning supplies and use them for titration? I don't know any way to get pure chemicals here, so my only source for strong chemicals is off-the-shelf caustic cleaning supplies from the hardware store.

They sell "90-100% oxalic acid dihydrate", "25-34% hydrochloric acid", "100% sodium carbonate", and an NaOH blend: "95-100% Sodium Hydroxide / 1-5% Sodium Carbonate / <1% Dipropylene Glycol Methyl Ether". Plus I can get the standard non-caustic brewing salts, and "80% lactic acid".

Any way to test these things against each other and come up with a strong acid and base of known concentration? The HCL seems promising, the NaOH... not as much.

edit: I found "98+% clean sodium hydroxide" from a soap-making store. Important extra fact: I can get 4.01 pH calibration solution. It's unmarked, but I believe that is always 1.02% potassium hydrogen phthalate, which is used to standardize HCL. Can I standardize the hardware store HCL with buffer solution, and then standardize the soap NaOH to the HCL?
 
Last edited:
Is there any way to standardize unknown strength cleaning supplies and use them for titration? I don't know any way to get pure chemicals here, so my only source for strong chemicals is off-the-shelf caustic cleaning supplies from the hardware store.

They sell "90-100% oxalic acid dihydrate", "25-34% hydrochloric acid", "100% sodium carbonate", and an NaOH blend: "95-100% Sodium Hydroxide / 1-5% Sodium Carbonate / <1% Dipropylene Glycol Methyl Ether". Plus I can get the standard non-caustic brewing salts, and "80% lactic acid".

Technically you could use anything but a strong acid and strong base will make life much easier. Of the choices HCl and NaOH blend will do.

Any way to test these things against each other and come up with a strong acid and base of known concentration?
Yes, if you have something of known "strength". Of what you have available it would seem that chalk from the home brew supplier is probably the most pure and not subject to pickup of water of hydration but it will adsorb some so it should be dried in an oven at above 100 °C before being weighed out.

edit: I found "98+% clean sodium hydroxide" from a soap-making store. Important extra fact: I can get 4.01 pH calibration solution. It's unmarked, but I believe that is always 1.02% potassium hydrogen phthalate, which is used to standardize HCL. Can I standardize the hardware store HCL with buffer solution, and then standardize the soap NaOH to the HCL?
Go with this NaOH. Not that a little sodium carbonate (which may well be what makes up the other 2% of this product is - it may also be CaCO3 as NaOH is deliquescent and picks up CO2 from the air).

In any case, start by drying some CaCO3 powder in the oven. Cool it (preferably in a dessicator) and weigh out 1 gram as accurately as you can. Suspend this in 100 mL (or so - doesn't really matter) of DI water which you have boiled and cooled while covered. Add the HCl in small increments until pH 5 is reached. At pH 5 the charge on bicarbonate and carbonate in the solution is -0.0400 mEq per mmol. In the CaCO3 powder the charge on the carbonate ion is -2.00 mEq/mmol. Thus -.04 - (-2.00) = 1.96 mEq/mmol of protons have been absorbed in getting to pH 5. You have suspended 1000/100.09 = 9.99101 mmol CaCO3 and thus used up 1.96*1000/100.09 = 19.5824 mEq. In addition to protonating CO3-- ions you have acidified the water to the extent of .001 mmol/L. Assuming you suspended the chalk in 0.1L of water that means an additional 0.0001 mEq protons which, of course, you can neglect but if you want to carry it for illustration purposes that means you have consumed a total of 18.5824 mEq of protons.
print 1000/100.09

37% HCl (23 Be') is about 12.39 N. Thus if you had acid of this strength you would use 18.5824/12.39 = 1.51076 mL. The acid you are looking at is somewhat weaker than 37% so you will find that you need a bit more but the point is that 1.5 - 2 mL isn't much and you would need to be able to measure it to 0.01 mL or so. 18.5824 (mEq)/1.51(mL) = 12.3062 (mEq/mL = N) but 18.5824/1.52 = 12.2253 N. Thus, especially as your goal is to get acid of 1 N or less, diluting the acid you buy with 10 or 11 parts of DI water before you begin. The mix will be about 1 N and that is a reasonable strength to work with in doing the malt titrations.

Now here's the real catch. When you react CO3-- with acid the reaction takes place pretty quickly at first (baling soda volcano) but then slows way down. When you start to add acid to the suspension it will fizz so be sure to do it slowly enough that it doesn't fizz over. If you get spatter on the sides of the beaker wash them back down into the liquid with a squirt of DI water from a wash bottle. Eventually the fizzing upon addition of more acid will slow, the suspension will turn clear and the addition of further acid still will result in a sharp drop in pH. Don't think you are finished at this point because if you come back an hour later the pH will have risen again and if you come back an hour later still it will have risen further. It doesn't really matter what the pH eventually arrives at as long as it is stable indicating that all the invisible CaCO3 microcrystals have reacted. IOW if it settles and pH 4.5 instead of 5 you don't have to despair and start over because you have to be at 5 and this is a good thing because you may not stabilize for 24 hours! Just use number for the charge on carbo and the charge on water calculated for the final pH. I haven't experimented with this enough to be able to give you fuller guidance but I suspect that keeping the pH near 4 may lead to faster equilibration.

Now once you have standardized HCl you can standardize an NaOH solution against it. The molecular weight of NaOH is 40 so to make a 1 N solution of it you would need to dissolve 40 grams in a liter of water. Work fast with this stuff as it picks up water from the air pretty fast. Make up your solution and put, say, 10 mL of it in a small beaker of flask. Add 7 mL f your approximately 1 N HCl solution and titrate until pH 7 is reached. At this point you have added as many protons as there were OH- ions in the 10 mL of NaOH solution. Suppose you found your HCl solution to be 0.876 N and that you added a total of 11.3 mL of it to 10 mL of NaOH solution to reach pH 7. This tells you that 10 mL of NaOH solution contains 11.3*0.876 = 9.8988 mEq OH- and is thus 0.98988 N.
 
Yes, if you have something of known "strength". Of what you have available it would seem that chalk from the home brew supplier is probably the most pure and not subject to pickup of water of hydration but it will adsorb some so it should be dried in an oven at above 100 °C before being weighed out.

Thanks, A.J.! I'll start buying toys soon, and probably try this out next month.
 
I was looking at the Riffe Part III paper today and marveling at the fact that he seems to think that using a 'fiduciary malt' is simpler than using standardized acids and bases. This approach introduces the extra work of having to measure out the fiduciary malt and can destroy the quality of any estimates made if the properties of the fiduciary malt are not precisely known and if the fiduciary malt is too similar to the test malt. But musing over this while looking at a series of posts in other threads where they are struggling to figure out how much sodium bicarbonate to add the light came on: sodium bicarbonate would make a fine fiduciary malt when working with dark grains. The procedure would follow what Riffe has in his paper except that in one of the jars one would put nothing but 1/20th of of a kg (50 g) of the malt (ground) to be tested. After adding 100 mL of water to that (mash thickness of 2L/kg) pHDI is measured. For the next jar we would like the pH to be say 0.05 higher. We know that these malts have buffering of about -50 mEq/kg•pH so we need to absorb about 50*0.05/20 = 0.125 mEq in the second jar in order to have its pH approximately 0.05 units higher than pHDI. The molecular weight of NaHCO3 is 84 so that means we'll need 84*50*0.05/20 = 10.5 mg. Thus we would put no bicarbonate in the first jar, 10.5 mg in the second, 21 in the third, 31.5 in the 4th and so on up to as many jars as we can handle. This, of course, brings up the question of how to measure out 10.5 mg of NaHCO3 accurately. Probably the best way is to accurately weight out 1 gram of it, put that in a 1L container (best if it is a 1 L volumetric flask but we don't have to be NBS precise here), dissolve it and make up to the mark with DI water. For 10.5 mg, measure out 10.5 mL if solution. Now you don't have to add exactly 10.5 mL. But you want to know, as accurately as possible, how much you added so if you over fill a graduated pipet or don't shut off the valve on a buret in time that's OK. Just record what you added. Clearly using liquid you would put that in the container you are going to use to measure the "mash" water first, then make up to 100 mL, then warm. Probably a good way to work is while observing a clock or timer. Prefill all the beakers or jars with the 50 gram ground malt samples. Add prewarmed water the first and place it in the water bath. During the net five minutes prepare the warm water with the first bicarbonate addition. At the end of the period add the water to the second malt sample and add it to the water bath. And so on. At the end of 30 minutes take the first sample out of the water bath and place it in the cooler. Five minutes later take out the second and so on. When each sample is cool measure and record its pH. Sample should cool at 5 minute intervals and that should be sufficient for stable pH readings.

When finished you will have a set of pH and bicarbonate addition pairs. For each pair, compute the number of protons absorbed per kg malt from 20*mg*QAcid(pH, 6.38, 10.38) - 1)/84. Values of QAcid(pH, 6.38, 10.38) + 1 can be had from the following table

pH..QAcid(pH, 6.38, 10.38) +1
3.5 0.99868348
3.55 0.99852308
3.60 0.99834316
3.65 0.99814137
3.70 0.99791506
3.75 0.99766125
3.80 0.99737663
3.85 0.99705747
3.90 0.99669962
3.95 0.9962984
4.90 0.99584861
4.05 0.99534442
4.10 0.99477932
4.15 0.99414603
4.20 0.99343643
4.25 0.99264144
4.30 0.99175097
4.35 0.99075374
4.40 0.98963721
4.45 0.98838744
4.50 0.98698893
4.55 0.98542448
4.60 0.98367502
4.65 0.98171948
4.70 0.97953457
4.75 0.9770946
4.80 0.97437134
4.85 0.9713338
4.90 0.96794809
4.95 0.96417725
5 .00 0.95998111
5.05 0.95531623
5.10 0.95013582
5.15 0.94438978
5.20 0.93802478
5.25 0.9309845
5.30 0.92320991
5.35 0.9146398
5.4 0 0.90521141
5.45 0.89486131
5.5 0.88352646
5.55 0.87114555
5.60 0.85766055
5.65 0.84301857
5.70 0.82717385
5.75 0.81009007
5.80 0.79174265
5.85 0.77212115
5.90 0.75123169
5.95 0.72909898
6.00 0.70576816

Tabulate mEq/lg vs pH and proceed as in the No.2 to determine a, b and c. It's a good idea to use the values obtained to plot the curve they describe and to plot the indidual measurement pairs on the same chart. The closer the points to the curve the more confidence you will have in your values.

I said in No. 7 that I didn't think you could use lactic acid to do a malt titration because it is a weak acid and it's strength varies with pH. But we can easily compute its strength from each measured pH and could use it if we could be confident that lactic acid labeled 88% were really 88%. It's the same with sodium bicarbonate - it is a weak base but we can be confident of its strength and AFAIK it does not pick up water from the air. Manufacturers specs seem to indicate that their products are more than 99% pure and that's what we want it for - so that we can be confident that the strength we calculate from the table is accurate. Do not, therefore, dry it it an oven before weighing it out. This will cause it to decompose (2NaHCO3 --> 2H2O + Na2CO3 + H2O). Its clearly a lot less expensive, easier to obtain and more stable than a standardized NaOH solution.

What's given above is clearly for dark malts with pHDI's outside of the desired mash pH range on the low side. What of the base malts? There we need an acid. I've rejected lactic on the basis of uncertainty about the labeled strength. Just as we did with bicarbonate we would hope to find a readily available pure powdered acid. Malic acid meets the requirements. I'll post a strength table for it in another post.
 
Last edited:
I like the idea of being able to do all 4 titration steps at the same time. And the ability to use lactic acid (I’d like to think a new bottle of lactic acid would be very close to 88%), malic and sulfuric acid. AJ thank you for sharing the bicarbonate titration steps and table of QAcid pairs.
 
You do NOT need the Solver to find a, b and c. Excel can invert matrices and that's all it needs to be able to do in order to solve this directly. It requires you to do some matrix operations and I won't get into how to do that here but assuming you know how here's the steps:
1) Put all your pH measurements in a column
2) In the column next to it put (pH - pHDI) for each pH measurement
3) In the column next to that put values for (pH - pHDI)^2
4) In the column next to that put values for (pH - pHDI)^3
The columns of steps 2,3 and 4 are to be interpreted as the Matrix A in the following steps. It has N rows, one for each measurement and 3 columns
5)Select an empty area of the spreadsheet of 3 rows and N columns
6)In the formula bar type = TRANSPOSE(range) with the range specified by dragging over the Matrix A area. Be sure to hit Ctrl-Shift Enter (not just enter). This puts the matrix A_T (the transpose of A) into the new area
7) Select an empty area of 3 rows and 3 columns.
8)Type MMULT(range1,range2) with range 1 being the range of A_T and range2 that of A into the formula bar. The order is important. This puts (A_T*A) into the new area.
9)Select a empty range of 3 rows and 3 columns
10)Type MINVERSE(range) with range being that of (A_T*A) into the formula bar. This puts (A_T*A)^-1 i.e. the inverse of (A_T*A) into the new area.
11)Select a clear area with 3 rows and N columns.
12)Type MMULT(range1,range2) with range1 being that of (A_T*A)^-1 and range2 that of A_T. That puts ((A_T*A)^-1)*A_T into the new area. This is the Moore-Penrose "pseudo inverse of A (only square matrices have inverses).
13)Select an area with 3 rows and 1 column
14)Type MMULT(range1, range2) into the formula bar. range1 is the range of ((A_T*A)^-1)*A_T and range2 the range of proton additions corresponding to the measured pH's. a, b, and c will appear in the new area.

That's it. Probably easier to use solver but once you have this set up you can paste new pH and Q vectors into it and the answers will update automatically.
 
i spent the afternoon writing a VBA function that finds a,b, and c using the Moore-Penrose solution described in #19 but all the matrices are hidden, You just put your data pairs in adjacent columns in the sheet, drag over 3 cells to reserve a place for a, b and c to be returned type the function name in the formula bar, drag over the data to tell the function where it is and hit ctl-shift enter and the answers appear instantly. The really nice thing about this is you can see what fiddling with a potentially bad point does right away. No need to go through the process of setting up another solver run.

People on the list will be getting a new version of the spreadsheet with this included.
 
This is in response to a request for tips on how to titrate a malt sample and determine a, b and c from the titration data. The request was that this be a Sticky. I'm not sure that is justified as I doubt many will want to undertake malt titration.

The procedure we are going to try to simplify for home brewing here is found in my paper "Predicting and Controlling Mash pH Using Simple Models for Mash Component Acid/Base Characteristics" MBAA TQ vol. 52, no. 1 2015 pp 3 - 12.

The simple model for malt is

∆Q(pHz) = a*(pHz - pHDI) + b*(pHz - pHDI)^2 + c*(pHz - pHDI)^3

∆Q(pHz) is the number of mEq of protons that must be added to a kg of malt mashed in a reasonable amount of DI water to bring the pH of the mash to pHz. If ∆Q(pHz) > 0 then one adds the protons by adding acid. If ∆Q(pHz) < 0 then protons must be removed, i.e. absorbed, by adding base.

pHDI is the distilled water mash pH for this malt. a, b, and c (called a1, a2 and a3 in the paper) are coefficients which describe, along with pHDI, the malts acid/base characteristics. To determine them we add acid and base in different amounts to small portions of the malt in question mashed with DI water and measure the pH. The first pH measurement is made with no acid or base addition and the resulting pH is pHDI. We then plot the acid or base additions vs the measured pH and find values of a, b, and c which produce a curve which best fits the data.

Some example data is shown in the plot below where each circle represents a measurement. It's based on the Weyermanns pilsner malt curve in the paper but has been played around with to illustrate a few points which we'll get to. For starters, you measurements won't come out on nice whole tenths of a pH. The job is to find the curve which best fits the data in the rmse sense. We'll have Excel do that for us but first let's find out how to get the measurements.View attachment 580983
What You Will Need

1)0.1 N Sulfuric Acid (Hach)
2)0.1 N Sodium Hydroxide (Hach)

We need to measure out protons and proton absorption capacity in mEq. An 0.1 N solution of sulfuric acid supplies 0.1 mEq/mL an an 0.1 N solution of sodium hydroxide absorbs 0.1 mEq/mL. Obviously you can get 0.1 N solutions from many places or make them up yourself if you know how and know how to standardize them. Hach sells them in 1 liter bottles and all the work is done for you but you know what the catch is - you have to pay for that.

3)Some way to measure small quantities of these solutions accurately. Syringes that can be read to tenths of a cc should do. A buret is better, an adjustable pipetter better still and a Hach Digital Titrator best (IMO)*.

4)A means to finely grind small quantities of malt. I use a Waring spice grinder.

5)A means to weigh out 50 grams of malt accurately

6)A source of DI water

7)A means to measure out 100 mL of water (great accuracy is not required here)

8)Beakers (ideally stainless steel) to hold the mini mashes (50 grams of powdered malt with 100 mL of water)

9)A good pH meter calibrated frequently

10)A means for holding the beakers at about 50° C.

11)pH 4 and 7 buffers,

Procedure:

1) Transfer small quantities of pH 4 and 7 buffer into containers which can be immersed in the water bath and place them there.
2) Measure out 100 mL DI water and place the container in the water bath at 50 °C
3) Finely grind somewhat over 50 grams of malt. Mix the grind thoroughly so the husks are evenly distributed throughout the flour. Measure 50 grams as accurately as possible and transfer to a beaker.
4) Place the beaker in the water bath (a weight will probably be necessary to keep it submerged)
5) Wait until the temperature of the DI water is 50 °C (or close to it)
6) Mix the water with the malt stirring thoroughly. Continue to stir periodically until 25 minutes have passed.
7) During the waiting period calibrate the pH meter using the 50 °C buffers.
8) Measure the pH (yes, at 50 °C). Be sure the reading is stable before accepting it (at least a couple of minutes)
9) Remove the beaker from the water bath and cool it to room temperature.
10) When the mash is at room temperature measure the pH again.

11) Look at the pHDI reading obtained at 50 °C. in the example data set it was 5.49. This is right in the middle of the mash pH range so you will be using the parameters to calculate ∆Q's at pH's above and below this value. Thus you will have to titrate with both acids and bases. It's good to have as many data points as possible, of course, but this is pretty tedious work so you might decide to shoot for measurements separated by 0.2 pH. You know that the buffering of most malts is about -40 mEq/kg•pH so that to raise the pH of a mash of this malt to 5.69 you would need to absorb something like 40*0.2 = 8 mEq/kg. You are mashing 50 gram (equals 1/20 kg) samples and so would require 8/20 = 0.4 mEq proton absorbing capacity. You have 0.1 N NaOH solution which absorbs 0.1 mEq/mL and thus would need 4 mL of that. Accurately measure 4 mL of 0.1 N NaOH into about 90 mL of DI water
12) Repeat from Steps 2 - 8 except instead of measuring out 100 mL of DI water in Step 2 make up the water NaOH mix from Step 11 to 100 mL.

From here on you continue taking data, plotting each point as it comes in, until your data set looks something like the set of circles in the picture above. To collect lower pH points use exactly the same reasoning as given in Step 11 but for acid instead of base. IOW add 4 mL of the 0.1 N H2SO4 to 90 mL of water and use that.

Obviously, as noted above, this is going to get tedious pretty quickly. About the only recommendation I have there is to see if you can handle more than one titration at a time. That way you get three points in an hour instead of only 1 but you have to figure out a protocol that lets you juggles three mashes at once. Preparing 3 malt samples and 3 water/acid/base mixes and then mixing water and malt staggered 3 - 4 minutes apart seems to work for me.

Note that your measured pH's will not be on nice even tenths of pH's nor will they be evenly spaced in 0.2 pH increments above and below pHDI. This is because, of course, the malt you are measuring won't have a = -40 mEq/kg (what a actually is is what we are trying to find out.

I'm going to stop here for fear of overflowing the size limitation. To be continued.

*This is in part because the digital titrator is accurate, easy to use, can be used for other things such as measuring alkalinity and hardness and because the NaOH titrant comes in a cartridge which is sealed in a plastic bag thus protecting it from air (CO2)
 
Thank you for the great post and sorry to come back to it now for some noob questions.

Why you suggest to calibrate and measure pH at 50 °C rather than 20°C after sample cooling? Is it to avoid mash pH shift with temperature?

Which is the reason behind making every time a new mash sample starting with a different amount of acid/base in the strike water rather than tritate a single sample with more acid/base addition?

Thank you!
 

Latest posts

Back
Top