At long last, ASTM D8045 is here!

Way back in 2002, some Brazilian chemists came up with a smart new method to measure the free fatty acid (FFA) content of vegetable oils using thermometric endpoint titrimetry (TET) (1). I tried it out, and it worked like a charm. It was fast, accurate, and very precise, all the things you need in process and quality control. The endpoints are highly reliable, and best of all, the sensor doesn't crud up like a potentiometric pH electrode, and require cleaning and rehydration every other titration or so. The probe can be stored dry, and is ready for use 24/7. I thought, this method is essentially about the determination of weakly acidic species in a non-aqueous environment. I had in mind a replacement for ASTM D664, the potentiometric titration procedure for Acid Number in lubricating and other mineral oils. For those of you who have had the misfortune to work with this purely dreadful procedure, you'll know what I mean. Let's face it, glass membrane pH probes just aren't meant to operate in a non-aqueous environment. The glass membrane has to be kept in a hydrated state, and the reference junction has to be kept clean. Operating in a dehydrating, oily medium isn't really their thing. Yes, I know that companies have spent years and a lot of money developing pH probes which do work a bit better in non-aqueous solutions, but really, they're a compromise at best. All hail to those clever Brazilians for coming up with a procedure which is really useful!

It was 2005 when I adapted and tried out the new TET method on both vegetable and mineral oils. When our company Multitrator sold the TET enabling technology to Metrohm, it became part of the technology package we handed over. Fast forward to early 2008, when I introduced the method to Metrohm USA..Fast forward again to July 2016 when ASTM finally approved the procedure for the determination of acidity in oils, with the title: ASTM D8045-16  Standard Test Method for Acid Number of Crude Oils and Petroleum Products by Catalytic Thermometric Titration. Along the way, I did a bit of tweaking around the edges (improving the solvent mixture, figuring out a way to automate the addition of the thermometric indicator/catalyst), but really, the principle is the same as that outlined in the original paper.

It would be nice to sit back and relax, but there's more work to done. Firstly this procedure needs to be adopted in industries which make and use edible fats and oils; after all, it was originally intended to analyse those materials. Secondly, there is a crying need in the analysis of lubricating oils to replace the potentiometric titration for the analysis of Base Number. The potentiometric method suffers the same sorts of problems evident in the Acid Number determination - most of which are caused by the probe. I worked up another TET method for Base Number using another thermometric indicator. The idea was taken from a paper published over 40 years ago (2). I'm hoping that this procedure can be approved by ASTM more rapidly than that for Acid Number!. It is also a procedure that has wider uses than just the analysis of lubricating oils. I have tested it out, and it is excellent for the determination of purity of weak organic bases in non-aqueous media. The titrant is exactly the same as used for the potentiometric procedure, and can be bought off the shelf.

References.
1. M. J. D. Carneiro, M. A. Feres Júnior, and O. E. S. Godinho. Determination of the acidity of oils using paraformaldehyde as a thermometric end-point indicator. J. Braz. Chem. Soc. 13 (5) 692-694 (2002)

2. E. J. Greenhow and L. E. Spencer (1973) Ionic polymerisation as a means of endpoint indication in non-aqueous thermometric titrimetry. Part 1. The determination of organic bases. Analyst, 98, 81-89

Rapid, Robust and Reliable - The New TET Way To Determine Acid Number In Crude Oils And Refinery Intermediates And Products

Good agreement with result obtained by conventional titrimetric method
Many crude oils contain a range of acidic substances which can cause corrosion damage to refineries during processing. These acidic substances can be categorized as “napthenic acids” which have been defined as an “unspecific mixture of several cyclopentyl and cyclohexyl carboxylic acids with molecular weight of 120 to well over 700 atomic mass units”. Sulfur-containing acids may also be present, and can contribute to corrosion problems. The bulk acid content of an oil is defined by its AN (Acid Number) value, defined as mg KOH/g of oil. The unit price of a crude oil can be discounted significantly according to its AN, so it is important to be able to analyze the AN value with accuracy and precision.

To date, the conventional method to determine the AN value of crude oils has been a potentiometric titration based on ASTM method D664. The success of this method is dependent on the type of crude oil analyzed, and the skill of the analyst in identifying problems which may affect the result. Further, the performance of the pH probe is affected by dehydration and fouling of the sensor glass membrane, as well as its reference junction. Potentiometric pH probes may require cleaning and regeneration (rehydration) of the sensor after only a few titrations. In contrast to the potentiometric method, the thermometric probe requires no conditioning and minimal maintenance. It may be stored dry between titrations, and is always ready for service. It is thus very suitable for use in remote locations with minimal laboratory facilities.

In thermometric endpoint titrations (TET), the endpoint is determined by the rate of change of temperature of the titration solution as the titrant is added. In the case of non-aqueous titrations of weakly acidic species, the temperature inflection at the endpoint is too low to be detected reliably. In this case, a thermometric indicator (paraformaldehyde) is employed. At the endpoint, the first trace of excess KOH titrant catalyzes the endothermic decomposition of the paraformaldehyde. This provides a reliable marker for the endpoint. The titration is carried out in an anhydrous mixture of xylene and 2-propanol. There is no need to add water to the solvent, as is the case with potentiometric titrations.

A sample of a very dark, highly viscous crude oil African origin (Doba, Chad) was provided for evaluation of the procedure.

Reagents:

1. Titrant: c(KOH) = 0.1 mol/L potassium hydroxide in 2-propanol

2. Solvent: 1:1 mixture of xylene with 2-propanol

3. Thermometric indicator: paraformaldehyde (e.g., Sigma Aldrich cat. no. 158127)

4. Standard: c(C₇H₆O₂) = 0.1 mol/L benzoic acid in 2-propanol
Method Outline:

Approximately 3 g of crude oil is weighed into a titration vessel, and 30 mL of solvent mixture added. Approximately 0.5 g of paraformaldehyde is then added. The titration is then commenced, and stopped after appearance of the endpoint inflection.

In the case of the semi-automated titrations reported on here, the paraformaldehyde is added as a powder. A level 1/8^th kitchen teaspoon measure was used to dispense the paraformaldehyde. In the case of fully automated titrations employing a sample changer, the paraformaldehyde is slurried with the solvent mixture, and delivered by peristaltic pump as part of the titration program. The titration can be stopped automatically after the endpoint appears.
Results:

Due to the small amount of oil provided, the average AN was computed from a limited number of determinations, with sample masses ranging from approximately 0.5 to 4 g.. Individual results were 4.1, 4.3, 4.2, 4.2, 4.2 and 4.3 with an average of 4.2±0.09 mg KOH/g. The result obtained appears to be insensitive to the amount of sample mass, at least within the range tested.

The customer recorded a value of 4.4 mg/100g KOH by a procedure similar to that specified in ASTM D664.

TET plot of acids in crude oil in 1:1 xylene/2-propanol, using c(KOH) = 0.1 mol/L as titrant.  

Rapid automated analysis of nitrating acid mixtures by TET

Mixtures of concentrated nitric and sulfuric acids are used in the preparation of explosives and propellants for military and civilian use. Tight control of the amount of the individual acids is necessary in order to obtain products of consistent composition and performance. Rapid feedback to the production department of analyses is essential to permit corrective action to out of specification nitrating mixtures without losing valuable time. Traditional manual analytical procedures may not be capable of delivering results within a timely manner, and so automated analytical procedures may be considered. Further, traditional manual methods may be subject to analyst error, and a high degree of training and supervision may be required in order to obtain results of sufficient accuracy and precision.

A new automated TET analytical protocol for the analysis of sulfuric, nitric and nitrous acids has been developed, employing the Metrohm 859 Titrotherm thermometric titration system. Nitrous acid is a by-product of the nitration process, and its presence has to be compensated, in order to obtain an accurate value for the nitric acid content. The protocol requires two separate titration sequences. In the first titration, the HNO2 content is determined by a direct TET, employing 0.1 mol/L KMnO4 as titrant. The result is saved by the software. In the second sequence, the sulfuric acid content of the sample is determined by TET with 1 mol/L BaCl2, and the result saved. The second titration of the sequence then starts. This is the determination of the Total Acid content, by TET with 2 mol/L NaOH. In the calculation at the end of the second titration, the HNO3 content is determined by subtracting the H2SO4 and HNO2 contents, after converting the results to an HNO3 equivalent.


A complete H2SO4-HNO3-HNO2 analysis can be completed in less than 7 minutes. For higher analytical productivity, a sample changer can be used.  The analyst only needs to weigh the samples into the titration vessels, and place them in the rack of the sample changer. All other steps are fully automated. The procedure is also suitable for completely automated on-line analysis.

Typical TET titration plots are illustrated in Figs. 1, 2 and 3. The solution temperature as a function of titrant delivered is shown as a red curve. The endpoint is accurately located from the inflection in the second derivative curve.

Fig.1. TET plot - titration of H2SO4 in nitrating mixture with 1 mol/L BaCl2

Fig. 2. TET plot - titration of Total Acids (H2SO4+HNO3+HNO2) with 2 mol/L NaOH

Fig. 3. TET plot - titration of HNO2 with 0.1 mol/L KMnO4

It should be emphasized that these three different titrations were all accomplished with the same sensor. This is simply a highly sensitive, rapidly responding electronic thermometer. It requires no calibration and minimal maintenance, and is thus ideally suited for routine process and quality control applications.

Science in my pocket.

It seems that I can be fairly accused of not keeping up with what's going on out there. A few months ago, I posted on the slowly developing trend to taking advantage of the advantages that smart mobile phones offered in being able to be used as highly portable scientific instruments. Recently, a friend alerted me to a Kickstarter funding project which is set to take this concept to a much higher level. An Israeli company called Consumer Physics has been working away on a miniature NIR (Near Infra Red) analyzer for the past couple of years, and has targeted the beginning of 2015 for release of their SCIO instrument. 

The ultimate idea is that you would point it at an object, and it will give you its chemical analysis. Of course, nothing in life is easy, and NIR is no exception. NIR has been around for decades, but the availability of cheap, powerful computers to run the software to process the complex signals coming from NIR instrumentation (and thereby yield useful results) has increased popularity in recent years. 

When NIR light is directed at an object, chemical components of that object reflect or absorb various wavelengths of that light in varying amounts. That reflected light comprises a spectrum, which is analyzed by a spectrometer contained in the handpiece. The signal from the spectrometer is sent via Bluetooth to the smart phone, where the software compares the spectrum to those analyzed previously from similar samples for which the chemical analysis is known. The accuracy of NIR analysis is dependent on a number of factors which include:

- the accuracy of the chemical analysis of the components contained in the calibration samples

- the condition of the calibration samples (how they were prepared and presented for NIR analysis)

- the similarity of the samples under test to those used for the calibration

- the condition and similarity of the surface of the object, compared to those used for the calibration

- sophistication and relevance of the algorithms used to deconstruct and quantify the spectra.

So the SCIO NIR instrument doesn't yet rival the "Tricorder" used to analyze absolutely everything in Star Trek episodes, but if the project is realized as conceived by the developer, it will (in my opinion) represent the greatest advance to date in the development of portable chemical analyzers.

Aside from the brilliance in condensing a laboratory-sized instrument to a device seeming no larger than modern electronic car "keys", to me, the stand out features are the complete liberation of the sensor and the inspired use of the Cloud to refine computation algorithms as new data is logged. No longer is the sensor attached to the signal conversion and display device by a restricting cable. It is now truly and completely wireless. Every dedicated analyzer with a sensor bound to it by a length of cable is now obsolete.

As I've said before, exciting times ahead!

- 

Smart pHone?

I got my iPhone 4 over three years ago. Yes, I know, by today's standards it's ancient, but I come from a generation which was taught to use things up and wear them out - it's still working nicely. I realized shortly after getting it I had a computer in my pocket, and since then I have been looking for sensors, adaptors and apps which could turn my phone into a portable scientific instrument.

With built-in location services and a clock, the time and place of the measurement can be absolutely pinpointed, and transmitted by email or SMS to remote locations. A photo of the measurement location can also be taken and transmitted. Just the thing for the roving chemist.

The trouble is, until fairly recently, there's been nothing available on offer. There are lots of handy apps for the analytical chemist, and I've got them on my phone. Periodic tables, molecular mass calculators, and regression analysis calculators for preparing calibration curves. They're good and they're handy, and they've come to my rescue when I've been at a customer's and couldn't boot up my notebook to the internet, but somehow, I felt a lack....of, well, something better...until I found the Sensorex PH-1 pH meter for Apple mobile products a couple of years ago. It was cool, and I had to have one. It plugs into the connector of my iPhone, and a pH probe with a BNC connector plugs into it. The software is free from iTunes, and it all works very nicely. There's multi-point calibration, and you can save the calibration. You can email the readings, and you can name the location of your reading either however you like or with its GPS location.

So it all works very nicely. There's a couple of  problems chief of which is the fact that the iPhone 5 and later models use a different connector. That is easily solved with an adaptor, but you'll need a Lightning > 18-pin adaptor. However, I'm still very happy with this little device.
It seems that Sensorex have now resolved this problem with their new SAM-1 Smart Aqua Meter, which now connects to smartphones (Apple and Android devices) through the headphone socket. The new meter also performs conductivity and ORP measurements as well. Reportedly, shipments start mid-March 2014 on this product. Another thing that I like is that you can buy online, and that PayPal is accepted as payment.

I note that Sensorex are promoting their "smart" probes, which are auto-recognized by the SAM-1. Great, but they're probably not the first in this area. Way back in 2000, our Multitrator thermometric titration system went to market with our Intelligent Probe thermometric sensors. The probes contained a chip which stored the probe's manufacturing data as well as the calibration data which converted mV to temperature. There was also the capability for storing your own calibration data if you wanted to do some really accurate enthalpimetric measurements.

There is a remaining niggle with the SAM-1. You still have to connect a probe directly to the meter. Wouldn't it be nice if you could be free of those tangly, interference-prone analog wire connections, and make your pH, conductivity, ORP (and maybe even ISE) measurements wirelessly and digitally?
It seems that there are moves in this direction. A company called Sensorcon has brought out a multi-sensor device called Sensordrone, which connects via Bluetooth to Android devices. 

Out of the box, it does all sorts of things ....

... and you can buy add-ons, such as a pH measurement module ....

as well as a dissolved oxygen sensor extension. The makers are promoting as a platform for third-party hardware devices. 

I can see interesting days ahead in the smartphone/portable lab. area. It's always fun to see what the future brings.

A salty tale ....

Way back in 2007, I wrote a piece on a simple chemical reaction which is now proving to be highly useful in chemical analysis for process and quality control. To save you having to scroll down to the bottom of the blog, here it is again:

You can use this reaction to determine by thermometric endpoint titration (TET) sodium, aluminium and fluoride (but not potassium), because the reaction is exothermic. I'm going to stick my neck out here, and say that this reaction offers the first viable titration method for analyzing sodium. Sure, there's been others, but none have proven to be successful. A titration based on the insolubility of zinc uranyl acetate has been around since 1931, and a complexometric procedure was reported in 1970, but my argument is that if they were of any real practical use, you would in fact see them being used. You don't. And yet titration is a great technique for routine process and quality control. Automated titration instruments don't take up much bench space, infrastructure is minimal, and lower-skilled operators can be used, providing they are well trained and well supervised.

In the case of TET, frequently it is unnecessary to filter sample solutions prior to titration. Further, the sensor is really just a fast-responding electronic thermometer. The probe doesn't need calibrating and requires little maintenance, the titrations take place in regular vessels (no insulation required), you don't need a reference electrode, the probe is always 24/7 ready - what could be easier? 

Thermoprobe thermometric titration sensor

"But wait a minute", I hear you say, "we're already titrating for sodium in food - we use silver nitrate as the titrant. That's all we need to do, isn't it?"

Well, no. It isn't good enough any more. The silver nitrate titration assumes that all sodium present is in the form of sodium chloride, common salt. You actually titrate chloride (and other halides) with the silver nitrate, and assume that all chloride is present as sodium chloride. The trouble is, that not all sodium in food comes from added sodium chloride. Lots of other sodium salts are added to processed foods; some are preservatives, some are flavour enhancers, and then there's emulsifiers and also stabilizers; all sorts of different sodium salts are used in foods. There's something else to consider: some manufacturers are seeking to reduce the sodium content of the food, while minimizing the impact of the "saltiness" of the food. They're doing this by substituting potassium chloride for some of the sodium chloride that's normally added. Look at this label from a container of margarine:
Potassium chloride is also being used as a sodium chloride replacement in snack foods as well. Check for the food code E508 on the label, if potassium chloride isn't mentioned by name.

The titration itself is quite straightforward. However, as in every analysis involving food, you have to pay attention to the sample preparation. The main aims are to liberate all the sodium from the food matrix, and to obtain a solution that is suitable for titration - not too viscous, and no lumps. Fortunately, you don't have to worry about turbidity or fine particulates, because there's no sensitive sensor membrane to foul or reference junction to clog. Even fatty foods can be handled with the appropriate sample preparattion. When you're considering the determination of sodium by TET in a food for the first time, there is a need to think carefully about the sample preparation options you need to consider.

Recently, I wrote a paper which gives some examples of the determination of sodium in various foods. You can download a PDF copy of the paper "Novel method for determination of sodium in foods by thermometric endpoint titrimetry (TET)" by clicking on the link. It was published in the January 2014 edition (Volume 3, Number 1B) of the open-access Journal of Agricultural Chemistry and Environment. If you don't have much time, there's a condensed version as a presentation-style PDF which can be downloaded from here. 

Goodbye to Gove (Part 2)

"Part 1" of my "Goodbye to Gove" dealt with one of the more significant technical challenges we faced in the laboratory during the early years of the Gove alumina refinery, which Rio Tinto Alcan is now in the process of closing down. This second part deals with the most critical challenge to the still young refinery, in which the lab. played an important part.

I've mentioned the word "fun" in Part One of these reminiscences of my technical life at Gove. Fortunately, there was more fun to come after we got the Bayer liquor analysis problem sorted out, which fortuitously put us in good shape for the next challenge. A few years after start-up, we were put in the position of changing the type of alumina that the company produced. If we didn't change the alumina product, the Gove refinery would go out of business.

A short explanation. Aluminium metal is smelted from alumina, Al2O3, That alumina is produced by decomposing aluminium hydroxide, Al(OH)3 at high temperatures.

In the illustration above, the red stuff is the bauxite, from which the aluminium hydroxide is refined (the white stuff). The white stuff is smelted to produce aluminium metal.

To produce aluminium, the alumina is put into a large carbon-lined electrolysis cell (a "pot"). The alumina is dissolved in molten cryolite, Na3AlF6, and lots of electricity is used to reduce the alumina to aluminium metal. You can read a description of the process here. The type of alumina the Gove refinery was designed to make was called "floury" alumina, which was a good descriptor. If you pinched a bit of this alumina between your fingers, it made little peaks, just like finely milled baking flour. It comprised mainly alpha-alumina (corundum), and was made by calcining Al(OH)3 along with a trace of AlF3 as a catalyst at very high temperatures, about 1200 deg. C.

There were process advantages in making floury alumina. Plants making floury alumina could be highly productive, making a lot of alumina per cubic metre of the recirculating aluminate liquor in the plant. There were manufacturing advantages to the smelters as well. However, there was one overwhelming disadvantage - these smelters were highly polluting. When the carbon anodes in the smelting pots burnt, the off-gases which also contained fluoride-containing gases and particulates polluted the surrounding countryside. Many of the smelters would scrub these off-gases with water, but they weren't very efficient. The solution was to employ a property of alumina to contain the fluoride in the smelting pots, and not release it into the environment.

If you don't calcine the alumina all the way through to corundum, you have a product which possesses a high adsorption affinity for all sorts of gases and gas-entrained fine particulates. The smelter design could be changed first to put lids on the pots, and then to lead the off-gases over the incoming alumina feed to the pots. Thus the potentially polluting fluoride could be recycled back to the pots, instead of escaping into the atmosphere. Modern aluminium smelters are highly efficient and essentially non-polluting. Trouble was, we weren't making the right kind of alumina. It wasn't just a case of calcining our product less, we would also have to make the particles coarse and able to withstand the rigours of the new dry gas scrubbing process. 

The scanning electron micrographs at left illustrate a coarse particle of agglomerated aluminium hydroxide, and below it, a particle of the coarse "sandy" alumina which is produced by calcining the aluminium hydroxide.
We didn't know how to make this product. Our parent company Alusuisse approached companies who did know how, but they held the technology very tightly. The technology could be purchased at a crippling price, but it would also result in a considerable reduction in the production capacity. The company faced a stark choice: either pay the exorbitant asking price and suffer a loss of production as well, or be eventually obliged to close down due to market pressure. Alusuisse did neither, thanks to their brilliant chemist and engineer, Dr Otto Tschamper. Dr Tschamper invented a process which held the promise of maintaining plant productivity while manufacturing a grade suitable for use in modern smelters. It was a massive improvement over the technology being used by our competitors. Dr Tschamper told me later that when he first announced the new technology at a technical meeting in the USA, he was laughed at. It's nice when you eventually get the last laugh, isn't it?

The technology had been successfully trialled in Europe, now the confirmatory work had to be done at Gove. Fortunately, our lab. was up to the challenge. Although Alusuisse wasn't originally keen on the idea, we had built up some considerable expertise in modelling plant processes at the lab. scale in order to service the demands of our process engineering colleagues of the technical department. We also had the new thermometric titration liquor analysis, which played a critical role in accurately measuring the mass balances required in evaluating the new process.

I was also fortunate to have a staff of excellent graduate chemists, among them Hans-Peter Breu, who would take a leading role in solving the many problems involved in adapting the new process to the Gove plant through clever experimental design and execution. Cut off from the mainstream of the then highly secretive world of alumina R&D, we had to teach ourselves. Firstly, we had to learn how to make coarse, strong gibbsite particles, then we had to learn how to do this in Gove plant aluminate liquor which was heavily contaminated with organic carbon compounds whose source was the Gove bauxite itself. The process which had worked in Europe didn't work with the organic-laden Gove liquor. It was an exciting time as we gradually learned how to make it work in laboratory-scale reactors. We had no access to the American laboratory test technology, so we had to learn everything from scratch. It wasn't just exciting, it was fun. 

I left Gove around the time of the successful pilot plant tests, firstly to work for Dr Tschamper in Zurich, then later to lead a laboratory supporting Alusuisse' alumina operations around the world in their research institute in Neuhausen, Switzerland. Hans-Peter Breu kept on developing his expertise in the highly successful new process, seeing it implemented and adapted in alumina refineries around the world. 

I heard the other day that after the refinery is shut down, the lab. staff will be reduced to just three, as activities at Gove will comprise just the mining and export of bauxite.