Driving with Biomass

The concept of using liquid biofuel dates back to the year 1895 when the German engineer Rudolf Diesel (1858–1913) developed the first engine running on vegetable oil. The motors of that time with their large injectors could easily cope with viscous vegetable fuels. However, due to the low petroleum prices, engine technology was increasingly tailored to consume low-viscosity conventional fuel. Consequently, vegetable oils were only sought after in times of high oil prices.
Not until vegetable oils were derivatised was low-viscosity biofuel available. In a reaction that is catalysed by a base, acid or enzyme, a vegetable oil or animal fat is reacted with methanol to yield fatty acid methyl esters (FAME, biodiesel) and glycerin as by-product. The latter, if produced in high quality, finds use as a valuable feedstock in the cosmetic and pharmaceutical industry. The above base-catalysed transesterification is considered as the most promising production process [1] (fig. 1). Due to the reversible character of the reaction, a large excess of alcohol shifts the equilibrium to the products side and thus ensures total conversion to the esters. After completion of the transesterification reaction, the biodiesel phase is separated from the more dense glycerin phase by gravitational settling or centrifugation. Subsequently the methyl esters, which still contain large amounts of residual alcohol, traces of dispersed glycerin and unreacted sodium hydroxide or soaps, are cleaned by a water-wash. Remaining water and poorly water-soluble impurities, such as the unreacted feedstock or the mono- and diglycerides, are removed by further steps such as distillation or stripping.

In 1908, some years after the development of the diesel engine, Henry Ford (1863–1947) designed the Ford Model T to run on ethanol. However, the low petroleum prices and the seemingly inexhaustible fossil fuel reserves also displaced the ethanol. Not until the worldwide oil crisis in 1973, Brazil and the US launched their first ethanol programs and thus paved the way for their actual leadership position concerning production and utilisation of bioethanol.
Generally, bioethanol is made from products containing sugar, starch or lignocellulosic biomass. The microbial fermentation of biomass-sourced sugars via yeast is a well established technology, applied commercially on a large scale. In contrast, starch biomass with its larger carbohydrates is not directly fermentable. Prior to yeast-induced fermentation, starch-containing feedstock has to be converted to sugars. Fermentation yields relatively dilute aqueous solutions of ethanol, which, for their later use as a fuel, are distilled to provide 95% ethanol. The “anhydrous” 99% ethanol is mainly produced via physical water absorption by molecular sieve.

Despite all the advantages mentioned at the beginning of this paper, biofuels had to struggle for acceptance right from the start. Reports highlighting engine problems due to poor quality biofuel discredited the promising biogenic route. Low-quality biodiesel, often produced from crude feedstocks in uncontrolled home-brewing plants, contained detrimental contaminants, resulting in injector fouling, enhanced corrosion and clogging of the fuel system. Not until reliable quality standards were defined did the quality of biofuels and thus the confidence of the consumer and the automobile industry improve. The major biodiesel standards, which commonly serve as reference for other standards, are the ASTM D 6751 from the American Society for Testing and Materials and the European EN 14214 (table 1a). Additionally, there exists the separate standard EN 14213 defining the minimal requirements for biodiesel used as heating oil or as a blending component for heating oil.

On the one hand these standards include fuel-inherent properties such as the oxidation stability or the iodine value. These so-called structure indices originally served to exclude the use of certain vegetable oils or animal fats as feedstocks [2]. On the other hand there are properties that are basically related to the production process. These parameters, also called quality indices, indicate the content of unreacted starting material in the biodiesel. Process-related parameters comprise the acid number as well as the glycerin, methanol, water and sodium hydroxide content. As mentioned later, the determination of both water content and acid number is crucial for the quality control of the feedstock and for optimising the production process [3–5].

While the standardisation of biodiesel in Europe has been well established by the EN 14214 since 2003, the European standard for bioethanol, the prEN 15376 is currently under approval. In contrast, the leading ethanol producers, the US and Brazil, adhere to two well-established standards, the ASTM D 4806 and the ASTM D 5798 for denatured fuel ethanol only and for mixtures of bioethanol and gasoline (Ed75-Ed85), respectively (table 1b).

In view of the fact that the quality control of biodiesel begins with the refining of the vegetable oil feed, this discussion stresses the importance of the acid number and the water content of the feedstock used.
However, emphasis lies on specifications and test methods prescribed by the two biodiesel standards, namely the determination of the oxidation stability, the iodine and acid number as well as the water, alkali metal and alkaline earth metal content. Titrimetric and ion chromatographic analyses referred to in the ASTM D 4806 bioethanol standard are adressed as well.

Oxidative Stability
As mentioned earlier, biodiesel is readily biodegradable, allowing its use in environmentally sensitive areas. However, this environmental advantage also means that the fuel is less stable, which affects storage behaviour. Especially derivatives of polyunsaturated fatty acids, such as linoleic (C18, 2 double bonds) and linolenic acid (C18, 3 double bonds) with one or two bis-allylic methylene positions, are highly susceptible to oxidation. During the first step of fuel oxidation (hydro)peroxides form through a free-radical chain mechanism. In the second step the radicals produce short-chain aldehydes, ketones and carboxylic acids (acid number increases). Under certain conditions, a radical-initiated polymerisation can form insoluble polymers, which in turn can clog fuel lines, filters and pumps. These drawbacks are less pronounced in unrefined vegetable oils containing natural antioxidants. During refining these antioxidants get partly lost and oxidation stability decreases. However, premature degradation can be overcome by the addition of synthetic antioxidants. Their effectiveness can accurately be investigated with the so-called Rancimat method.

The Rancimat method mimics the oxidation of a biodiesel sample at a fixed temperature, usually far above ambient. The result is then extrapolated to the stability under real-world conditions. In practice, a stream of purified air is passed through the heated sample (usually 110 °C) and is subsequently bubbled through a vessel containing deionised water (fig. 2). The resulting oxidation products – volatile organic acids, predominantly formic acid – are swept from the sample into the water, thus increasing its continually monitored conductivity. The point at which the maximum change of the oxidation rate occurs is the so-called induction time. The PC software evaluates the induction time automatically from the maximum of the second derivative of the conductivity with respect to time.

Experimental
In order to determine the temperature dependence of the induction time, sample amounts between 3 and 6 g biodiesel were analysed at 100, 110 and 120 °C (fig. 3).

The results (table 2) agree with the Arrhenius equation, according to which a temperature reduction of 10° C should result in an approximate doubling of the induction time. At 110 °C the investigated biodiesel sample has an induction time of 6.3 h. It thus complies with the minimal requirements of EN 14112 in EN 14214 (6 h) and ASTM 6751 (3 h).

Iodine Value in Biodiesel
The iodine value (IV) or iodine number is another stability index and a measure for the unsaturation in organic compounds. It is the amount of iodine in g that can be added to 100 g of the sample and is used as an indicator of the number of double bonds. The higher the IV, the higher the number of double bonds. Originally, the IV in EN 14214 had the function of excluding certain feedstocks for biodiesel production. However, since the IV does not consider the positions of the double bonds within the compound, it does not correlate well with the oxidation stability. Knothe et al. [6] showed that different free fatty acid (FFA) structures can give the same IV. Consequently, the IV is increasingly understood as a rough indicator. Stability specifications known as API (allylic position equivalents), BAPE (bis-allylic position equivalents) and above all the previously described Rancimat test characterise the oxidation stability of biodiesel much more accurately.

Experimental
After the titer determination, 0.15 g biodiesel sample is dissolved in 20 mL glacial acetic acid and treated with 25 mL Wijs solution as iodinating reagent, consisting of iodine monochloride in glacial acetic acid. After 5 minutes, 15 mL potassium iodide solution is added. As in classical iodometry, the excess of iodine is titrated with standardised 0.01 mol/L sodium thiosulfate solution. A Pt Titrode is used for endpoint indication.

The investigated biodiesel sample has an IV of 114.4 (table 3 and fig. 4) and thus meets the requirements of EN 14214 with a permitted maximal value of 120 g iodine per 100 g sample.

Acid Number in Biodiesel and Acidity in Bioethanol
High fuel acidity is associated with corrosion and engine deposits, particularly in the fuel injectors. The acid number (AN) or acid value of edible oils or their corresponding esters indicates the quantity of fatty acids and mineral acids (negligible) present in the sample. According to ASTM D 664 and EN 14104, the AN is expressed in mg KOH required to neutralise 1 g of FAME. The acidity of bioethanol is contained in both ASTM D 4806 and ASTM D 5798 using the method ASTM D 1613. It covers the determination of total acidity as acetic acid.
The AN is included in EN 14214 and ASTM D 6751, which suggest the methods
EN 14104 and the ASTM D 664, respectively. Both standards stipulate a non-aqueous potentiometric acid-base titration and limit the acid content to 0.5 mg KOH per g sample. Alternatively, ASTM D 974 can be used for coloured samples; it involves the non-aqueous colorimetric titration using KOH in isopropanol as the titrant and p-naphtholbenzein as the indicator).

Besides the quality control of biodiesel, the AN plays a significant role in the quality control of feedstocks. Generally, the glycerides should have an AN below 1 mg KOH/g [3, 7]. Higher ANs lower the ester yields and increase NaOH consumption for neutralisation. Feedstocks containing high levels of fatty acids should therefore preferably be processed to biodiesel via an acid-catalysed transesterification.
Additionally, increasing ANs, when compared to the initial AN of the biodiesel, can point to ongoing fuel degradation or the intrusion of water (hydrolysis of the FFAs).
In the following, the determination of the AN of a biodiesel sample is illustrated using method EN 14104.

Experimental
Between 14 and 15 g biodiesel sample is dissolved in 50 mL bioethanol/diethyl ether mixture (1:1, v/v). The sample is titrated potentiometrically with alcoholic potassium hydroxide. After each titration the Solvotrode, a pH glass electrode that has been especially developed for non-aqueous acid-base titrations, is thoroughly rinsed with isopropyl alcohol. The regeneration of the membrane is achieved by immersing the electrode in water for at least three minutes.

The determined AN of the biodiesel sample is 0.202 mg KOH/g (table 4). This value complies with the requirements of ASTM D 6751 and EN 14214, which both stipulate a maximum AN of 0.5 mg KOH/g.

Water Determination in Biodiesel and Ethanol
In the biodiesel production process, water contamination of biodiesel plays a significant role in both the quality control of the feedstock and the end product.
Biodiesel, although considered to be hydrophobic, can contain as much as 1500 ppm of dissolved water, excluding suspended water droplets. The presence of water in biofuels reduces the calorific value, enhances corrosion, promotes the growth of microorganisms and increases the probability that oxidation products are formed during long-term storage. Additionally, water cleaves the ester bond of the FAMEs via hydrolytic degradation. The same applies for the glycerides in the feedstock. The liberated FFAs consume the added NaOH, forming soaps and emulsions that increase viscosity and seriously hinder the phase separation of glycerin. Because of this, all materials used in the biodiesel production process should be essentially anhydrous.

Several methods exist for the determination of water: loss on drying, reaction with calcium hydride, Karl Fischer titration (KFT), Fourier Transform Infrared (FTIR) and Raman spectroscopy as well as dielectric measurements. Among these, KFT is certainly the method of choice when trace amounts of free, emulsified or dissolved water have to be accurately determined in a reasonable time.
KFT is based on the stoichiometric reaction of water with iodine and sulfur dioxide in the presence of a short-chain alcohol (R` = CH₃, C₂H₅) and an organic base (RN), according to the following equation:

R`OH + SO<sub>2</sub> + 3 RN + I<sub>2</sub> + H<sub>2</sub>O → 3 RNH<sup>+</sup> + R`OSO₃^– + 2 I^–.

Whereas volumetric KFT is applied to samples with water contents ranging from approximately 1 up to 100%, the coulometric technique is ideally suited for low water contents in the range of a few µg/g. In the volumetric KF technique a titrant containing iodine is directly added to the sample via a buret. In contrast, in coulometric KFT iodine is generated electrochemically from iodide directly in the titration cell. In both cases iodine reacts with the water in the sample.
ISO 12937 in EN 14214 prescribes coulometric KFT for the determination of the water content. According to EN ISO 12937, the test results must meet the following requirements regarding repeatability:

The difference between two test results, obtained by the same person under identical test conditions, may exceed the following value r for the repeatability only once in 20 cases:

r=0.01874√x,

where x is the mean value of all test results given as a mass fraction in percent rounded off to 0.001%.

By means of direct coulometric titration using different commercially available KF reagents, the water content of a biodiesel sample is determined and the repeatability r calculated.

Experimental
Between 0.9 and 3 g biodiesel sample is directly injected into the reaction solution with a syringe. Once all the available water has reacted (equivalence point), the indicator electrode detects the first excess of iodine and the KFT stops. The amount of water is calculated by measuring the electric charge needed for iodine generation.

Irrespective of the KF reagent used, all results are in the same ppm range (table 5). The differences x_max-x_min are much smaller than the repeatabilities r defined by EN ISO 12937. This clearly shows that direct KFT provides a far better repeatability than is required by EN ISO 12937. The same applies for the automated pipetting system of Metrohm, which has been especially developed for high sample throughputs [8]. Accordingly, for ethanol the ASTM standard E 1064 in ASTM D 4806 prescribes the coulometric KFT of low water contents. For water contents > 2% the recommended test method is volumetric titration as per ASTM E 203 in ASTM D 4806.

Chloride and Sulfate in Ethanol
Contamination of ethanol with inorganic anions such as chlorides and sulfates can affect the engine performance because precipitating salts clog filters and fuel injector nozzles. Furthermore, these salts induce corrosion in the vehicle components in contact with the fuel. Against this background, the ethanol specification ASTM D 4806 limits the sulfate and chloride content to 4 and 40 ppm, respectively.

ASTM D 512 and ASTM D 7318 prescibe the use of potentiometric titration for chloride and sulfate, respectively. ASTM D 7319 presents a direct-injection suppressed ion chromatographic method for the determination of both anions.

a) Titration
As an example for potentiometric biofuel titrations, the determination of sulfate in ethanol according to ASTM D 7318 is presented. The determination of chloride, which can be carried out by mercurimetric or argentometric titration or with a Cl-selective electrode, is beyond the scope of this article.
100 g bioethanol sample is spiked with known amounts of a sulfate standard. After the addition of 1 mL 0.1 mol/L perchloric acid, the sulfate is precipitated with a lead nitrate solution. The Pb-selective electrode detects the first excess of lead ions at the equivalence point. A double junction Ag/AgCl (fig. 5a) or a glassy carbon rod electrode (fig. 5b) is used as reference electrode.

While sulfate concentrations between 5 and 10 ppm result in recovery rates of 100.7…105.6%, sulfate contents of 1 and 20 ppm provide recovery rates of 98.0…106.5% and 108.3…108.8%, respectively (table 6). Correlation coefficients of real concentrations versus determined concentrations for the double junction Ag/AgCl and the glassy carbon rod electrode are 0.9993 and 0.9991, respectively.

b) Ion chromatography
After direct injection, chloride and sulfate are separated on an anion exchange column and then determined quantitatively by suppressed conductivity detection (fig. 6).

The limits of detection for chloride and sulfate are 0.6 and 0.2 ppm, respectively. Even after 1500 ethanol injections containing denaturants and hydrogen peroxide, the analytical unit still provides stable retention times, repeatable peak areas and consistent concentration values [9]. This highlights the extraordinary ruggedness of the applied micro packed tri-chamber suppressor “MSM II” in long-term use.
The presented direct-injection IC system is 100% solvent compatible and ensures the accurate and precise determination of sulfate, chloride and other anions in full compliance with ASTM D 4806.

Alkali Metals and Alkaline Earth Metals in Biodiesel
After esterification and subsequent treatment, alkali metals and alkaline earth metals may be present in biodiesel as unwanted residues. Standard DIN EN 14 214 permits a cumulative concentration of 5 mg/kg for both the alkali metals sodium and potassium and also for the alkaline earth metals magnesium and calcium. Both groups of cations can be determined rapidly and accurately in a single ion chromatographic run (fig. 7).

Experimental
The samples are extracted with dilute nitric acid, dialysed and then injected directly into the IC system. The complete sample preparation procedure and analysis takes place fully automatically. The combination of instruments used consists of the 861 Advanced Compact IC with Metrohm inline extraction and dialysis.

Antioxidants in Biodiesel
As already mentioned, the oxidation stability of biodiesel can be improved by the addition of antioxidants. The addition of Baynox® to the biodiesel sample inhibits both the oxidation to corrosive acids and the formation of insoluble polymers. Although not regulated by standards, these substances are determined within the context of quality monitoring and for determining the amounts of additives to be added.

Experimental
Because of their structural similarities, vitamin E (α-tocopherol) and Baynox® can be determined together in a single analysis (fig. 8). To improve solubility, dichloromethane is added to the eluent and analyte solutions. The biodiesel samples should be diluted 1:1000. The analytes are separated at 35 C° and then determined quantitatively using UV detection.

Conclusions
This paper provides an overview of state-of-the-art analysis methods that facilitate fuel testing procedures and simultaneously help to reduce the danger of producing out-of-spec fuel. However, standards, especially in the biofuel environment, are subject to frequent changes. While standards will continuously be reviewed and updated (see the recent adoption of the Rancimat method EN 14112 in ASTM D 6751), interesting new analytical techniques such as the ion chromatographic detection of glycerin in biodiesel, the voltammetric determination of copper in ethanol or the measurement of the pH in ethanol will emerge.

References
[1] Blume A.M. and Hearn A.K.: Biofuels – Plant and Technology, 20–23 (2007)
[2] Knothe: JAOCS 83 (10), 823–833 (2006)
[3] Ma F. and Hanna M.A.: Bioresource Technology 70, 1–15 (1999)
[4] Ma F. et al.: Transactions of the ASAE 41, 1261–1264 (1998)
[5] Knothe G. et al.: The Biodiesel Handbook, AOCS Press, 304 pages, 2005
[6] Knothe K.: JAOCS 79 (9), 847–854 (2002)
[7] Bradshaw G.B. and Meuly W.C.: Preparation of detergents, US Patent 2, 360–844 (1944).
[8] Schlink R. and Faas B.: Water content determination in biodiesel according to EN ISO 12937, Pittcon 2007, www. metrohm.com/infocenter/posters
[9] Gandhi J.: Determination of sulfate in denatured ethyl alcohol by direct-injection ion chromatography and sequential suppressed conductivity, American Laboratory (2007)

A modified version of this article was first published in the PTQ (Petroleum Technology Quarterly) supplement: Biofuels – Plant and Technology, 23–30 (2007)

Authors:
Alfred Steinbach, Uwe Loyall, Barbara Zumbrägel, Christian Haider, Günther Spinnler, Regina Schlink and Andrea Wille