SEG-A-081

Quality Control According to DIN EN 14103, 14105, 14110

Application Book Volume 4 Quality Control Application Book Volume 4 Biodiesel Quality Control Contents

I. Determination of in biodiesel 5

II. Fatty acid methyl esters (FAME) Determination of total ester content and linolenic acid methyl ester contents 9

III. Determination of and glycerides 14 Part 1: Introduction and sample preparation 15 Part 2: Instrument configuration 17 Part 3: Optimization of the gas chromatographic method 20 Part 4: Signal identification 22 Part 5: Evaluation and quantification 25

Application Book Volume 4 Biodiesel Quality Control

Publisher: Shimadzu Europa GmbH Albert-Hahn-Str. 6-10 · 47269 Duisburg, Germany Telephone: +49 (0) 203 7687-0 Telefax: +49 (0) 203 766625 Email: [email protected] Internet: www.shimadzu.eu

Editorial Office: Uta Steeger · Phone: +49-203-7687-410 Ralf Weber, Angela Bähren

Design and Production: m/e brand communication GmbH GWA , Düsseldorf

© Copyright: Shimadzu Europa GmbH, Duisburg. Reprint, in whole or in part, is subject to the permission of the editorial office.

Windows is a Trademark of Microsoft Corporation. 5 I. METHANOL Biodiesel quality control according to DIN EN 14110

Determination of methanol in biodiesel

A

1 1 H2C – O – CO – R H2C – O – H H3CO – CO – R P A PB 2 CH3OH 2 HC – O – CO – R HC – O – H H3CO – CO – R B

3 3 H2C – O – CO – R H2C – O – H H3CO – CO – R

Triglyceride Glycerine FAME (Biodiesel)

Figure 1: Equilibrium reaction of the transesterification of plant oils A B () to fatty acid methyl esters (biodiesel)

iodiesel production has Figure 2: Equilibrium between the liquid- and gas-phase during headspace increased considerably in sample preparation. Although component A is present at a lower concentration B Europe in recent years. in the liquid phase, the concentration ratio in the gas-phase may shift, depend- In 1996, production was nearly ing on the vapor pressure of the individual components. Correspondingly, the 0.5 million tonnes. Today, the chromatograms resulting from headspace injection usually produce different production capacity is much signal ratios when compared with those resulting from liquid sample injection. higher. This development has also been supported by the automo- bile industry. In fact, several mil- The DIN EN 14214 standard ages (% m/m). DIN EN 14110 lion passenger cars and commer- specifies all legal limits for possi- describes the required method for cial vehicles have been approved ble by-products. For safety-relat- methanol determination. by their manufacturers to run on ed reasons and in order to com- System with headspace biodiesel fuel or higher biodiesel ply with requirements of the GC with headspace autosampler for high precision blends with mineral diesel. Com- automobile industry, the legal injection results pliance with the European Bio- limit for methanol residue has diesel standard DIN EN 14214 been defined as 0.2 mass percent- Biodiesel can consist of up to 100 Requirements of DIN EN 14110 ensures a continuously high qual- different fatty acid methyl esters, exceeded ity level. depending on the plant oils used during production. For the deter- The production of biodiesel is mination of methanol in such a based on transesterification of complex matrix, gas chromatog- plant oils or animal to fatty raphy in combination with acid methyl esters (FAME, Fig- headspace injection is the recom- ure 1). At a temperature of mended method. This enables approximately 60 °C, the plant separation of low-boiling point oil, consisting mainly of glycerin methanol from the high-boiling esters, is esterified using point fatty acid methyl esters methanol and a catalyst (alkali already during the sample prepa- hydroxide or alkali alcoholate). ration step. In order to quantitatively shift the equilibrium reaction towards The biodiesel samples are heated the biodiesel or FAME side, in a gas-tight 20 mL vial. Metha- methanol is added in excess. nol is a low-boiling point com- pound and is enriched in the After the reaction is finished, gas-phase while the high-boiling the excess methanol together point compounds remain in the with several by-products must liquid phase (Figure 2). After a be removed. Figure 4: GC-2014 and HT200H specific time, equilibrium is

6 Biodiesel quality control according to DIN EN 14110 I. METHANOL

reached and a constant volume can be sampled from the gas- 4.0 5.0 3 uV (x 100,000) Y = aX + b phase (headspace) of the vial and a = 4,601507e-002 b = 0.0 injected into the GC system. 3.5 4.0 R^2 = 0.9999938 2-Propanol R = 0.9999969 Calibration with minimum 3.0 3.0 complexity

2.5 Since the biodiesel matrix can 2.0

influence the equilibrium Concentration Ratio 2.0 2 between the methanol concentra- 1.0 tions in the liquid- and gas-phase, 1.5 calibration standards must be 0.0

prepared in methanol-free Methanol 0.0 2.5 5.0 7.5 10.0 1.0 biodiesel. The DIN EN 14110 Area Ratio biodiesel standard recommends preparation of a calibration curve 0.5 containing 0.01, 0.1 and 0.5 % Figure 5: Calibration curve of methanol in biodiesel in accordance with DIN EN (m/m) methanol in FAME. An 0.0 14110 using internal standard calibration internal standard should always be used when applying manual 0.0 2.5 5.0 7.5 Sample 1 2 3 headspace injection. When using a headspace autosampler, an Minutes Average concentration % (m/m) 0.009 0.100 0.500 internal standard will not be Reproducibility standard 0.00005 0.00100 0.00485 necessary but is, however, recom- deviation % (m/m) mended to increase the reliability Figure 3: Chromatogram of 0.01 % Reproducibility % RSD 0.54 0.99 0.97 of the results. The additional (m/m) methanol in biodiesel. 2-pro- work is kept within reasonable panol was added as internal standard. Table 1: Biodiesel sample spiked with three different methanol concentrations. limits: to each standard and Table 1 shows the reproducibility over 6 measurements of the same sample each biodiesel sample, 5 µL of using the same method on the same instrument. 2-propanol is added (Figure 3). a given time. The autosampler prepares several samples simulta- All measurements are carried out neously to ensure the highest At a methanol concentration of content were spiked with variable on a Shimadzu GC-2014AFsc possible sample throughput and 0.01 % (m/m) a signal-to-noise amounts of methanol. system in combination with the to prevent hold-up at the gas ratio of higher than 2000 :1 was HT200H headspace autosampler chromatograph. A gas volume of obtained. Thus, the technique Samples with the lowest concen- (Figure 4). Both instruments are 500 µL is sampled from the offers many options for improve- trations showed the highest dif- easy to operate and designed for headspace of each vial and inject- ments via application of larger ference of two out of six meas- daily routine analysis. ed into the GC-2014AFsc system split ratios or faster chromato- urement values 0.00012 % (m/m). with a split ratio of 1:10. graphic separation. DIN EN 14110 allows for For the calibration, 5 g of the 0.00151 % (m/m), a ten-fold respective standard solution as Additional parameters are recom- Results exceeding the higher deviation. well as 5 µL of 2-propanol are mended according to DIN EN standard requirements added to a 20 mL headspace 14110: The combination of the HT200H crimp vial. Corresponding to 5 g Temperature split injector: The calibration curve in Figure 5 and GC-2014AFsc offers an effi- of biodiesel sample, 5 µL of 150 °C demonstrates that the combina- cient as well as accurate automa- 2-propanol is added as internal Column oven (isothermal): tion of the HT200H autosampler tion solution for the routine anal- standard. The HT200H can han- 50 °C and the GC-2014 results in high ysis for methanol determination dle up to 40 sample vials per tray. Flame ionization detector: linearity over the concentration in biodiesel. Smaller scale bio- 150 °C range recommended in DIN EN diesel producers will profit from Automated sample Carrier gas (helium) linear 14110. Reproducibility of the this robust analytical system with preparation velocity: 35 cm/s measurement values for biodiesel its excellent price/performance Separation column used for samples is higher than the ratio. Sample preparation is fully auto- the measurements: requirements according to DIN mated, whereby each sample is Restek, Stabilwax-DA 30 m, EN 14110 (Table 1). Biodiesel shaken and heated to 60 °C for 0.32 mm, df 1.0 µm samples with very low methanol

7 9 II. FAME Biodiesel quality control according to DIN EN 14103

Fatty acid methyl esters (FAME)

Figure 1: Structural formula for linolenic acid methyl ester Determination of total ester content and linolenic (methyl cis,cis,cis-9,12,15-octadecatrienoate)

n the last years, biofuels have linolenic acid methyl ester is also fatty acid methyl esters in the become well established as determined. While this compound range from C4:0 up to C24:1. I an independent source of is a welcome essential acid in sal- energy in the industrial countries. ad oils, it is an interfering reactive The chromatogram in Figure 3 The percentage of biofuels in the compound in biodiesel that might shows the complete separation energy mix is expected to in- influence long-term stability of except for C4:0 and the oleic acid crease considerably within the biodiesel. methyl ester C18:1 cis/trans iso- next 10 years in order to reduce mers. The chromatogram time dependency on fossil fuels and to Instrumentation and can be reduced drastically since maintain energy prices at long- measurement a complete separation of all fatty term affordable levels. acids methyl esters is not re- A Shimadzu GC-2010AF with quired for the evaluation accord- Determination of iodine Based on the amount of freight AOC-20i autosampler (Figure 2) ing to DIN EN 14103. Good sep- value transportation, the need for diesel was used for FAME analysis. The aration is required for linolenic fuels and consequently biodiesel, GC is equipped with a Split/ acid methyl ester and the internal Requirements of DIN EN 14103 is significant compared to other Splitless injector (SPL) as well as standard C17:0 which is added to exceeded biofuels in the European Union. a Flame Ionisation Detector each biodiesel sample. All other Should the European Parliament (FID). Control of the GC-2010 fatty acid methyl esters from GC system prepared for agree with the proposal of the and AOC-20i autosampler as well C14:0 to C24:1 are evaluated as future challenges European Commission, EU as data acquisition was performed sum area. Therefore, separation member states will add at least by GCsolution software. Alterna- of single peaks is not needed. 10 % biofuel to fossil fuels before tively the AOC-5000 sampler can the year 2020. In Europe, the be used if the system will also be Problems could occur if animal demand for biodiesel will further used for quality control according fats are used as feedstock for increase, leading to a continuous to DIN EN 14410 (headspace biodiesel production. In contrast search for suitable plant oils and injection technique). to plant oils which consist of - animal fats that can be used as ty acids with even chain length raw materials. Experience shows Using the special Restek FAME- numbers, animal fats include also that the quality of the raw mate- WAX column it is possible to uneven numbers including C17:0. rial has an enormous impact on completely separate almost all Thus co-elution with the added the quality of biodiesel. Continu- fatty acid methyl esters from internal standard amount must be ous quality control therefore C6:0 to C24:1. For identification corrected for if necessary. remains necessary in order to of the individual fatty acid meet the demands of modern methyl esters the Supelco stan- Another problem might occur diesel generators. dard “37 Component FAME with plant oils containing fatty Mix” (Cat. No. 47885-U) was acids with chain lengths higher According to DIN EN 14103 used. The standard contains 37 than C24. The DIN EN 14103- biodiesel must consist of 90 % (m/m) or more of fatty acid methyl esters within the range of C14:0 up to C24:1. The designa- Figure 2: GC-2010AF with tion Cx:y refers to the carbon AOC-20i+s number or C-number (x) of the corresponding fatty acid as well as the number of double bonds (y). When y is larger than 0, it is an unsaturated fatty acid.

Accordingly, the polyunsaturated linolenic acid (Figure 1) is desig- nated as C18:3. The content of

10 Biodiesel quality control according to DIN EN 14103 II. FAME

acid methyl ester contents

2003 determines that all peaks Figure 4 on Page 12 shows the µV (x10,000) within the chromatogram range measurement of a biodiesel sam- from C14:0 to C 2 4 :1 are evalu- ple using the method described. 1.75 ated for fatty acid methyl ester For evaluation, the sum of all sig- content C but peaks outside this nal areas between C14:0 and 1.50 range must not be included in the C24:1 is calculated, taking also 1.25 evaluation (for calculation of C into account the areas of uniden- see equation on the right). Thus tified signals. Using the following 1.00 biodiesel produced from plant equation, the fatty acid methyl oils containing a considerable ester content C in % (m/m) can 0.75

amount of fatty acids C25:0 or be calculated according to DIN 0.50 higher might fail in quality con- EN 14103 as: trol according to DIN EN 14103. 0.25

0.00 Gas chromatographic ∑ ( A)-AISTD CISTD x VISTD 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 method C = x x 100 % AISTD m Minutes GC: GC-2010AF Autosampler: AOC-20i ∑A total signal area of all Injection vol.: 1 µL methyl esters from Figure 3: Supelco standard “37 Component FAME Mix” separated on a Restek Column: Restek C14:0 up to C24:1 FAMEWAX column. Not separated are the cis/trans isomers C18:1 (16). C4:0 not FAMEWAX (15734000 µV*s) separated from the solvent and the signals (1)-(6) representing the fatty acid 30 m, AISTD signal area of the internal methyl esters C6:0 up to C13:0 are not required for the quality control according ID 0.25 mm, standard heptadecanoic to DIN EN 14103. It was measured with the described gas chromatographic df 0.1 µm acid methyl ester C17:0 method except for the split ratio was 1:20. The naming of the individual compo- (Cat. No.12497) (2644200 µV*s) nents is given in Table 3 on Page 13. Carrier gas: Helium CISTD concentration of the Control mode: Constant heptadecanoic acid methyl linear velocity ester in the internal Injection mode: Split 1:50 standard solution used Linear velocity: 35 cm/s (constant) (10 mg/mL) Septum purge: 3 mL/min VISTD volume of the internal AL signal area of linoleic acid Injector temp.: 250 °C standard solution added methyl ester C18:3 (5 mL) (991200 µV*s) Oven temperature program m mass of the weighed ∑A total signal area of all FID settings: biodiesel sample (250 mg) methyl esters from C14:0 to C24:1 (15734000 µV*s) Rate Temp. Hold time For the biodiesel sample analyzed AISTD signal area of the internal (°C / min) (°C) (min) (Figure 4), a fatty acid methyl standard heptadecanoic 150.0 1.00 ester content of 99.0 % was acid methyl ester C17:0 5.00 240.0 6.00 determined. A fatty acid methyl (2644200 µV*s) Equilibrium time: 1.0 min ester content of 90 % or higher is required. The linolenic acid The result is 7.6 % linolenic acid Temperature: 250 °C methyl ester content L is calcu- which is well within the accept- H2 flow: 40 mL/min lated using the following equa- able range of 1 % up to 15 % Air flow: 400 mL/min tion as: (m/m). ‘ Makeup flow: 30 mL/min (He) AL Sampling rate: 40 ms L = ∑ x 100 % ( A)-AISTD Filter time const.: 200 ms

11 II. FAME Biodiesel quality control according to DIN EN 14103

µV (x10,000) Compound % (mm) Factor g iod / 100 g C16:1 0.20 0.95 0.2 4.5 C18:1 64.1 0.86 55.1 4.0 C18:2 18.7 1.732 32.4 3.5 C18:3 7.55 2.616 19.8 C20:1 1.32 0.785 1.0 3.0 C22:1 0.40 0.723 0.3 2.5 Calculated iodine value: 108.8 2.0 Table 1: Calculation of the iodine value for the sample analyzed (Figure 4). 1.5 The factors for the respective unsaturated fatty acid methyl esters are taken 1.0 from DIN EN 14214 Annex B. The mass percentages are calculated from the 0.5 chromatogram in Figure 4.

0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 Minutes iodine/100 g was determined Summary (Table 1, the limiting value is 120 g iodine/100 g). The price of biodiesel depends Figure 4: Biodiesel sample measured with the gas chromatographic method very much on the availability of described. For detailed information on the compounds see Table 3. When the same biodiesel sample cheap feedstock for the produc- is measured repeatedly with the tion process. This forces the pro- same instrument, DIN EN 14103 ducers to also use used oils or According to DIN EN 14214 allows a maximum difference in animal fats instead of fresh plant Annex B the iodine value can also methyl ester content of 1.6 % oil. Keeping up with this process be calculated from the methyl (m/m) for two arbitrary results. will be the future challenge of ester pattern obtained. However, Using the GC-2010 in combina- quality control for biodiesel. in case of a doubt DIN EN 14111 tion with the AOC-20i autosam- Especially the importance of the is the preferred method. For cal- pler, 24 measurements resulted in quality control according to DIN culation according to DIN EN a range over all results of 0.2 % EN 14103 will grow with this 14214 Annex B a good separation (m/m) – the difference between development. of the compounds in Table 3 the largest and the smallest meas- marked with asterisk is required. ured value was therefore 8 times The performance of GC-2010 For these compounds a factor to less than the required value in the in combination with AOC-20i – calculate the iodine value is given. norm. or alternatively with AOC-5000 – not only meets today’s require- The percentage by mass obtained For the determination of ments but is also prepared for from the chromatogram is then linolenic acid, a similar result is future challenges. The system is used to calculate the iodine value obtained. Over 24 measurements able to separate almost the whole of the sample, being the sum of the range over all results was composition of fatty acids, the individual contributions of only 0.01 % (m/m). including those with odd carbon each methyl ester, obtained by numbers from animal fat, in a multiplying the methyl ester per- The norm allows for a value ten reasonable chromatogram time. centage by its respective factor. times higher i.e. 0.1 % (m/m). This is the requirement for the All statistical results of the per- correct determination of the For the biodiesel sample in Fig- formed measurement sequence iodine value. ure 4 an iodine value of 109 g are summarized in Table 2. The precision of the system was Measuring value Methyl ester Linoleic acid content shown to exceed the require- content ments in the DIN EN 14103 by Mean value % (m/m) 98.9 7.55 a factor of ten. Standard deviation % (m/m) 0.06 0.06 Maximum value % (m/m) 99.0 7.56 Minimum value % (m/m) 98.8 7.55 Difference maximum/minimum 0.2 0.01 value % (m/m)

Table 2: Reproducibility over 24 measurements of one biodiesel sample. The table shows the statistical data including the relative standard deviation and the range of all results (max.-min. value).

12 Biodiesel quality control according to DIN EN 14103 II. FAME

Peak No. Compound Alternative naming Saturation Molecular formula CAS No.

01 Methyl hexanoate Methyl caproate, Hexanoic acid methyl ester, C6:0 C7H14O2 106-70-7 Methyl capronate

02 Methyl octanoate Octanoic acid methyl ester, Caprylic acid methyl ester C8:0 C9H18O2 111-11-5

03 Methyl decanoate Decanoic acid methyl ester, Capric acid methyl ester, C10:0 C11H22O2 110-42-9 Caprate

04 Methyl undecanoate Undecanoic acid methyl ester C11:0 C12H24O2 1731-86-8

05 Methyl dodecanoate Dodecanoic acid methyl ester, Lauric acid methyl ester, C12:0 C13H26O2 111-82-0 Methyl laurate

06 Metyl tridecanoate Tridecanoic acid methyl ester C13:0 C14H28O2 1731-88-0

07 Methyl tetradecanoate Methyl myristate, Tetradecanoic acid methyl ester, C14:0 C15H30O2 124-10-7 Myristic acid methyl ester

08 Myristoleic acid methyl ester 9-tetradecenoic acid, methyl ester, Methyl cis-9- C14:1 C15H28O2 56219-06-8 tetradecenoate, Methyl myristoleate

09 Methyl pentadecanoate Pentadecanoic acid methyl ester C15:0 C16H32O2 7162-64-1

10 Methyl cis-10-pentadecenoate cis-10-Pentadecenoic acid methyl ester C15:1 C16H30O2 90176-52-6

11 Methyl palmitate Hexadecanoic acid, Methyl ester C16:0 C17H34O2 112-39-0

12 Methyl palmitoleate* Methyl cis-9-Hexadecenoate, Palmitoleic acid C16:1 C17H34O2 1120-25-8 methyl ester

13 methyl heptadecanoate Heptadecanoic acid methyl ester C17:0 C18H36O2 1731-92-6

14 cis-10-heptadecenoic acid methyl ester Methyl cis-10-heptadecenoate C17:1 C18H34O2 75190-82-8

15 Methyl stearate Methyl octadecanoate C18:0 C19H38O2 112-61-8

16 Methyl oleate* Methyl cis-9-oleic methyl ester, Methyl cis-9-octa- C18:1 cis C19H36O2 112-62-9 decanoate (Z)-9-Octadecanoic acid methyl ester

17 trans-9-elaidic methyl ester* (E)-Methyl 9-octadecenoate C18:1 trans C19H36O2 2462-84-2

18 Methyl linoleate* Linolenic acid methyl ester; Methyl linoleate, C18:2 C19H34O2 112-63-0 cis,cis-Octadeca-9,12-dienoic acid methyl ester

19 Linolelaidic acid methyl ester* trans, trans-octadeca-9,12-dienoic acid methyl ester C18:2 C19H34O2 2566-97-4

20 ␥-linolenic acid methyl ester* Methyl (Z,Z,Z)-6,9,12-octadecatrienoate, C18:3 C19H32O2 16326-32-2 6,9,12-Octadecatrienoic acid methyl ester

21 Methyl linolenate* Linolenic acid methyl ester, C18:3 C19H32O2 301-00-8 Methyl cis,cis,cis-9,12,15-octadecatrienoate

22 Methyl arachidate Eicosanoic acid methyl ester C20:0 C21H42O2 1120-28-1

23 Methyl eicosenoate* Methyl cis-11-eicosenoate, cis-11-Eicosenoic acid C20:1 C21H40O2 2390-09-2 methyl ester

24 cis-11,14-Eicosadienoic acid methyl ester C20:2 C21H38O2 2463-07-7

25 cis-11,14,17-Ecosatrienoic acid methyl ester C20:3 C21H36O2 55682-88-7

26 cis-8,11,14-Eicosatrienoic acid methyl ester Dihomo-␥ -linolenic acid methyl ester, Methyl-DGLA C20:3 C21H36O2 21061-10-9

27 Methyl arachidonate Arachidonic acid methyl ester, C20:4 C21H34O2 2566-89-4 Methyl cis-5,8,11,14-eicosatetraenoate

28 Methyl 5,8,11,14,17-eicosapentaenoate cis-5,8,11,14,17-Eicosapentaenoic acid methyl ester C20:5 C21H32O2 2734-47-6

29 Methyl heneicosanoate Heneicosanoic acid methyl ester C21:0 C22H44O2 6064-90-0

30 Methyl behenate Docosanoic acid methyl ester C22:0 C23H46O2 929-77-1

31 Methyl erucate* cis-13-Docosen, Methyl cis-13-docosenoate; C22:1 C23H44O2 1120-34-9 Erucic acid methyl ester

32 cis-13,16-Docosadienoic acid methyl ester C22:2 C23H42O2 61012-47-3

33 Methyl tricosanoate Tricosanoic acid methyl ester C23:0 C24H48O2 2433-97-8

34 Methyl lignocerate Methyl tetracosanoate, Lignoceric acid methyl ester C24:0 C25H50O2 2442-49-1

35 cis-4,7,10,13,16,19-Docosahexaenoic Methyl cis-4,7,10,13,16,19-docosahexenoate, C22:6 C23H34O2 301-01-9 acid methyl DHA methyl ester

36 Methyl nervonate Nervonic acid methyl ester; Selacholeic acid C24:1 C25H48O2 2733-88-2 methyl ester, Methyl cis-15-tetracosenoate

Table 3: Components in the chromatograms in Figure 3 and 4. For the substances marked with an asterisk, the iodine value can be calculated with the given factor.

13 14 Biodiesel quality control according to DIN EN 14105 III. PART 1

Introduction and sample preparation

Property fatty acid in % Rape seed Sun flower Palm tallow Beef H C – O – CO – R1 H C – O – H H CO – CO – R1 Palmitic 5 6 42 28 2 2 3 Stearic 1 4 5 19 2 CH3OH 2 HC – O – CO – R HC – O – H H3CO – CO – R Oleic 60 28 41 45 3 3 Linoleic/Linolenic 30 61 10 5 H2C – O – CO – R H2C – O – H H3CO – CO – R

Table 1: Approximate content of fatty acids in different oils and fats Glycerine FAME (Biodiesel)

Figure 1: Transterification of triglycerols to FAME

iodiesel is an interesting Biodiesel mainly consists of tain amount of mono-, di-, and alternative to the decreas- methyl esters of the following triglycerides remains in the B ing resources of mineral fatty acids: biodiesel. Additionally, the sur- fuels. It can be manufactured plus of methanol and the by- from all kinds of plant oils or Palmitic acid C16 (saturated) product glycerine must be even fats from fryers. Stearic acid C18 (saturated) removed during the biodiesel Oleic acid C18 (unsaturated) production. In quality control Because of the high viscosity Linoleic acid C18 (poly unsatu- according to DIN EN 14105 or Control of transesterification direct use of those regenerative rated) ASTM D 6584-00 the contents of process energy resources in diesel cars the glycerine, mono-, di-, and requires expensive modifications The distribution changes depend- triglycerides are determined. Simplified standard to engine and tank. Therefore, the ing on the plant oil used (Table The methanol content is meas- preparation transesterification (Figure 1) of 1). In principle, all fatty acids ured with headspace technique the glycerol esters to fatty acid with chain lengths from C14 to described in DIN EN 14110. Improved reproducibility methyl esters (FAME) is pre- C24 may be found (in case of ferred. FAMEs are less viscous plant oils only the even num- Sample preparation and remain fluid even at low tem- bers). peratures. They can be used in Gas chromatography is a relative many diesel cars instead of min- FAMEs are the products of the technique. Every component eral diesel without changes to the reactions in Figure 1. must be calibrated before it can car. Presently the biggest market be quantified in unknown sam- for biodiesel is the blending with Dependent on process parameters ples. In the case of biodiesel one mineral fuels (e.g. B5; up to 5 % the transesterification is more or mono-, di-, and triglyceride is biodiesel in mineral diesel). less quantitative. Generally, a cer- calibrated, representative for ‘

No. Compound Alternative naming CAS No. Comments 1 MSTFA N-Methyl-N-(trimethylsilyl)-trifluoracetamide 24589-78-4 Derivatization of OH groups 2 Pyridine Azine 110-86-1 Buffer 3 Heptane 142-82-5 Solvent 4 1,2,4-Butanetriol 3068-00-6 Internal Standard 1 5 1,2,3-Tridecanolyglycerol Tricaprin 621-71-6 Internal Standard 2 6 Glycerin 56-81-5 7 1-Monooleoglycerine 1-Mono[cis-9-octadecenoyl]-rac-glycerol 111-03-5 Keep refrigerated at -20 °C (Monoolein) (monoolein) 8 1,3-Dioleoylglycerine 1,3-Di[cis-9-octadecenoyl]glycerol 2465-32-9 Keep refrigerated at -20 °C () (diolein) 9 1,2,3-Trioleoylglycerine 1,2,3-Tri[cis-9-octadecenoyl]glycerol 122-32-7 Storage below 8 °C (Triglyceride) (triolein)

Table 2: List of components required for biodiesel applications. Possible supplier: Sigma-Aldrich (Fluka). Due to derivatization with MSTFA all components must be water-free.

15 III. PART 1 Biodiesel quality control according to DIN EN 14105

the other components of each levels add the given volumes of group. Table 2 gives an overview stock solution A and C (Table 3) of chemicals needed for prepara- into 10 mL flasks. tion of standard and unknown samples. Required is a four level Derivatization of the calibration calibration of the components 6 mixtures to 9. Glycerin (6) is calibrated relative to the internal standard 1 Derivatization of the samples (ISTD1). The glycerides (7-9) are with MSTFA is crucial for Calibration level 1 2 3 4 calibrated relative to internal improvement of peak shape and Volume of stock solution A given into 100 µL 100 µL 100 µL 100 µL standard 2 (ISTD2). chromatographic separation. a 10 mL flask Add the given volume of stock 10 µL 40 µL 70 µL 100 µL Preparation of standard 100 µL MSTFA are added to each solution C into the same flask solutions of the four flasks prepared according to Table 3. Close the Table 3: Preparation of the four standard solutions In deviation from DIN EN 14105, flasks, tighten and shake them we recommend the following well. At room temperature the standard preparation. Three stock derivatization requires a mini- Calibration level Glycerin Monoolein Diolein Triolein solutions A, B and C are pre- mum of 15 minutes. After that conc. in mass % pared. Later, only the stock solu- time add 8 mL of n-heptane to 1 0.005 0.2 0.05 0.05 tions A and C are needed for each flask and shake again. The 2 0.02 0.8 0.2 0.2 preparation of standard levels and sample is now ready for measure- 3 0.035 1.4 0.35 0.35 unknown samples. ment. 4 0.05 2.0 0.5 0.5 Stock solution A: Table 4 shows the concentrations Table 4: Compound concentrations in the four calibration levels Internal standard in mass-% for the given compo- (given in mass %) 50 mg 1,2,4-butanetriol and nents in the respective calibration 400 mg tricaprin are given into a levels. 50 mL flask. Pyridine is added up to the 50 mL marking. Preparation and derivatization of unknown biodiesel samples Stock solution B: Glycerin dilution 100 mg biodiesel is given into a 50 mg glycerin are given into a 10 mL flask. 100 µL of stock 10 mL flask. Pyridine is added up solution A (internal standard) are to the 10 mL marking. added to the flask. Then add 100 µL of MSTFA, close the flask, Stock solution C: tighten, and shake it well. The Olein standard solution with derivatization takes at least 15 glycerin minutes at room temperature. 200 mg monoolein, 50 mg diolein After that time add 8 mL of and 50 mg triolein are given into n-heptane to each flask and shake a 10 mL flask. 1 mL of the stock again. solution B is added. Pyridine is added up to the 10 mL marking. Important

The three stock solutions can be Derivatized standards or samples kept in a refrigerator for weeks. must be measured within 24 For preparation of the standard hours.

16 Biodiesel quality control according to DIN EN 14105 III. PART 2

Figure 1: Mounting of retention gap kit with SGE metal connector Instrument configuration

ith this application the an exception being the Restek tor (e.g. SGE retention gap kit “Simple on-column” injection internal standard Vu2 connector with a maximum art. no. 052296). technique W method is used for operation temperature of 400 °C. quantification so an auto sampler For best performance both col- Significantly simplified column is not required, but recommend- Another alternative is the “simple umns should meet in the middle mounting ed for better reproducibility of on-column” injection technique of the connector. Since columns retention times. offered by Shimadzu. Instead of with ID of 0.25 mm or smaller fit More precise results through retention gap a special glass liner inside the ID 0.53 mm retention glass liner The injection technique is “cool is used. Design of the liner is gap, the optimum position is not on-column” injection (OCI) to similar to a glass connector (see easy to find. A compromise could Improved reliability avoid discrimination effects. With Figure 2). The injection is done be pushing the main column up OCI injection a thin needle of 26 into a small liner compartment to 5 mm inside the retention gap gauge – outer diameter (OD) 0.47 directly above the column. (Figure 1). More than 5 mm mm – is inserted into the column would lead to peak tailing and and the sample is injected com- Column mounting with poor results. pletely on the column. For retention gap accommodation of the syringe For accuracy and precision of needle, the capillary column must If a retention gap is used the con- the results it is essential that the have a minimum internal diame- nection to the main column is connection of the retention gap ter (ID) of 0.53 mm. In principle, often done with a metal connec- to the main column is gas tight. ‘ a 0.53 mm column can be used as main column for this application, but for better separation and faster chromatogram times a smaller ID is preferred.

Therefore, a retention gap – usu- ally a 2 m uncoated fused silica capillary with an ID of 0.53 mm – is used as pre-column. Connec- tion of the retention gap with the main column can be done with a metal connector (Figure 1). The preferred glass connectors cannot be used with this application due to the high oven temperatures – most glass connectors have a maximum temperature of 350 °C, Column and retention gap connection with metal connector

17 III. PART 2 Biodiesel quality control according to DIN EN 14105

Due to the high oven temperature biodiesel samples. If the first 20 - range the metal connector may 30 cm of the retention gap are cut eventually leak during operation. from time to time the main col- Therefore, it needs to be checked umn is less affected by contami- regularly. Because of this uncer- nants. tainty it is not recommended to use metal connectors in combina- With the liner the main column is tion with hydrogen as carrier gas. more affected by these contami- nants and it might be necessary Column mounting with liner to cut 20 - 30 cm from the col- umn regularly. However, the liner An alternative offers the “simple can be recycled. Best is to clean it on-column” injection technique in an ultrasonic bath using first offered by Shimadzu. The task of heptane and then methanol as the retention gap is taken by a solvents. It can be dried using a glass liner inside the injector. Any nitrogen gas stream or alterna- column ID from 0.1 to 0.53 mm tively, by heating the liner in an can be directly connected to the oven. It is important that it does liner. The column sticks inside not come into contact with con- the liner reduction and is addi- taminated surfaces. Deactivation tionally tightened with a graphite of the glass liner is not necessary. ferrule on the injector capillary adapter. Capillary column

This simplifies the column For good performance with this mounting significantly, and gas application the GC oven is oper- tightness even with hydrogen as ated at temperatures of 360 °C or carrier gas poses no problem any- higher. Unfortunately, most metal Figure 2: Mounting of column with simple on-column liner more. Two different liners are columns or aluminium coated available from Shimadzu depend- columns show severe problems ing on the column inner diameter. with recovery of di- and triglyc- erides. Generally with fused silica • 221-49381-02 for column high-temperature columns, the ID ≥ 0.32 mm recoveries are better. • 980-00371 for column ID ≤ 0.25 mm Often used is the HT5 capillary column from SGE. It’s a fused The injection is performed in a silica column with temperature small liner compartment above limits 380/400 °C – meaning the the column so a special on-col- column can be operated at 380 °C umn syringe is not required. A for longer times, at 400 °C only normal 23 gauge needle can be short term. Also possible is the used. HT8 phase from SGE but due to temperature limit 360/370 °C the Measuring the glycerol content in chromatogram time is extended. biodiesel, more precise results were found using the liner com- An alternative to the SGE col- pared to retention gap solution. umns could be the new Restek Figure 3: Cross-section through a capillary column high-temperature columns. Disadvantage of the liner is the Restek offers an RTX-Biodiesel lower capacity for contamination TG with a temperature limit of coming from biodiesel samples. 380 °C (Cat. No. 10294). Upon Retention gaps are good traps for request they also offer Shimadzu high-molecular components in a Fast GC column of the same

18 Biodiesel quality control according to DIN EN 14105 III. PART 2

uV (x 10,000) uV (x 10,000) 5.0 5.0

4.0 4.0 Monoolein Diglyceride

3.0 3.0 ISTD2 Tricaprin Butanetriol ISTD1

2.0 2.0 Triglyceride Glycerin

1.0 1.0

0.0 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 Minutes Minutes

Figure 4: Column conditioning. Between red and green line the column was Figure 5: Baseline drift or “ghost peak” due to liner contamination cycled about 20 times.

type, which reduces the analysis ning a baseline drift of about 30 tion, or the chemicals and sol- time to below 17 min. Restek also mV or more was observed. After vents used, which were not offers metal columns for biodie- 12 hours it was reduced by 10 to water-free. sel analysis e.g. MXT Biodiesel 20 mV. Figure 4 shows some con- TG (Cat. No. 70289). Due to a ditioning steps with a 0.1 µm film Normally, it is not possible to special deactivation process the column. Of course, a 0.25 µm eliminate these “ghost peaks” by metal columns show a good film column might show a more several injections of heptane. recovery for all biodiesel compo- severe baseline increase. A lower A new or cleaned liner has to be nents including the triglycerides. baseline drift of about 5 - 15 mV installed and 10 cm should be cut The MTX column is a mega-bore was observed with the Restek from the capillary column. column with an Integra-Gap, Rtx-5 high-temperature column. trapping contaminants before they enter the separation column. Ghost peaks

Conditioning of the system Even if the conditioning chro- matogram looks good it is recom- Before starting with measure- mended to inject pure heptane ments the conditioning of the once before measuring biodiesel system is essential. A new col- samples. Sometimes a baseline umn should be conditioned for effect might occur as shown in 12 hours using the temperature Figure 4. program for injector and detec- tor. A metal connector should be Most probably the liner was con- checked for leakage after the first taminated by contact with sur- two conditioning runs. faces covered by a fatty film (for example the human skin). If the Figure 4 shows a typical trend effect occurs during operation of of a baseline over more than 20 the system the reason may be the blank runs. samples themselves. For example the biodiesel additives, if A temperature program was used biodiesel from a gasoline station heating with 10 °C/min from was measured, overaged MSTFA 65 °C up to 380 °C. At the begin- which was used for derivatiza-

19 III. PART 3 Biodiesel quality control according to DIN EN 14105

Optimization of the gas chromatographic method

fter leak check and con- Oven temperature program ditioning of the instru- A ment a first sample can Rate (°C/min) Temperature (°C) Hold time (min) be injected. DIN EN 14105 65.0 1.00 recommends a non-polar column 15.00 170.0 0.00 with length of 10 m, ID (inner 8.00 270.0 0.00 diameter) 0.32 mm, film thickness 15.00 390.0 9.00 0.1 µm. Table 2: Equilibrium time: 0.5 min

Optimization of the method This is a good column to start according to columns used with, but most operators prefer Autosampler The next step is the optimization a longer column (12 to 25 m). of the method according to the A good separation is crucial With smaller ID a better resolu- Heptane is used as solvent for column used. The first step is tion can be achieved even with a biodiesel samples. The lifetime of injection of the standard solution shorter column (e.g. 12 m HT5, micro-syringes with biodiesel can 4 (highest concentration). ID 0.22 mm, 0.1 µm). In general, be extended if the syringe is a smaller ID allows a faster chro- rinsed regularly with a polar The standard contains one matography, but optimizing the solvent (e.g. methanol) besides representative component for gas chromatographic method heptane. With AOC-20i+s this each group of glycerides. might be more time-consuming. can be automated by using the three vial wash option. If no Butanetriol and tricaprin are Gas chromatographic method AOC-20s is used optional three added as internal standards. Even wash stations can be inserted into when refrigerated, the stock solu- GC: GC-2010AF with OCI the AOC-20i long rack. The tions for the standards alter slow- Autosampler: AOC-20i wash cycle could be 1 x heptane, ly. This is the reason for the addi- Column: HT5, 25 m, ID 0.32 1 x methanol and 1 x heptane tional peaks in the chromatogram mm, df 0.1 µm with retention gap again. (Figure 1). ID 0.53 mm Carrier gas: Helium AOC-20i settings Since the glycerin concentration Injection mode: Direct Injection volume: 1 µL counts significantly in the final Linear velocity: 50 cm/s (constant) # Rinses solvent post-run: 2 evaluation, a good separation Septum purge: 3 mL/min or alternatively: from the solvent peak is crucial # Rinses solvent post-run: 3 (Figure 1: more than 1 minute OCI temperature program (if three vial wash option is used) between solvent peak and glyc- # Rinses with sample: 2 erin). Rate (°C/min) Temperature (°C) Hold time (min) Plunger speed (suction): Low 60.0 1.00 Plunger speed (injection): High Typically the baseline drift 20.00 380.0 3.00 Syringe insertion speed: High becomes steeper if the oven tem- Table 1: OCI advanced settings fan: 100 °C Injection mode: Normal perature rises above 340 °C (Fig- Pumping times: 5 ure 1: chromatogram time 23 - 26 FID settings Temperature: 400 °C Used naming Alternative naming Glycerin H2 flow: 47 mL/min Air flow: 400 mL/min Butanetriol (ISTD1) Make-up gas flow: 30 mL/min Monoolein 1-Mono[cis-9-octadecenoyl]-rac-glycerol (monoolein) (He) Tricaprin (ISTD2) 1,2,3-Tridecanolyglycerol Sampling rate: 40 ms Diglyceride 1,3-Di[cis-9-octadecenoyl]glycerol (diolein) Filter time constant: 200 ms Triglyceride 1,2,3-Tri[cis-9-octadecenoyl]glycerol (triolein)

Table 3: Naming of biodiesel components

20 Biodiesel quality control according to DIN EN 14105 III. PART 3

uV (x 10,000) uV (x 10,000) 5.0 5.0 Column Temp. 375.0 Column Temp. 4.5 (Monitor) 375.0

Glycerin 4.5 (Monitor)

Monoolein 350.0

Diglyceride 350.0 4.0 Monoolein 4.0 Tricaprin ISTD 2 Tricaprin 325.0 325.0 Tricaprin ISTD 2 Tricaprin Butanetriol ISTD 1

3.5 Butanetriol ISTD 1 300.0 3.5 300.0 Diglyceride

Triglyceride 275.0 3.0 3.0 275.0 250.0 2.5 250.0 °C 2.5 °C 225.0 225.0 2.0 2.0 200.0 200.0 1.5 175.0 1.5 175.0 Glycerin

150.0 Triglyceride 1.0 1.0 150.0 125.0 125.0 0.5 0.5 100.0 100.0 0.0 75.0 0.0 75.0 -0.5 50.0 -0.5 50.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 Minutes Minutes

Figure 1: Chromatogram standard solution 4 Figure 2: Chromatogram standard solution 1

min). Fortunately, the drift is not (red marked value in Table 4). injected to check the detection too large here. Therefore, it is A combination of both is also limits. acceptable that the possible. In any case the hold elute within this drift. A bigger time X in line 4 has to be adjust- In the high-temperature range the drift could affect the integration ed so that the chromatogram time FID noise is 40 µV. Peak height accuracy of the whole group of does not become too long but the of the triglyceride peak is 2002 diglycerides eluting from 23 - 24 triglyceride signals are still µV. Thus, the signal-to-noise minutes. In such cases an increase recorded. ratio is 50. This is acceptable for of the linear velocity (up to 70 a 0.32 mm column, but could be cm/s) or changing of the oven Extending the temperature pro- improved by accelerating the temperature might help. gram has a disadvantage: the tri- chromatography for example glycerides remain too long on the using a higher linear velocity. The oven temperature program column. The peak might disap- can be slowed down by changing pear completely or become very the end temperature (Table 4, small. blue marked temperature). After optimization of the temper- Alternatively, the temperature ature program the standard solu- rate in line 4 can be decreased tion 1 (lowest concentration) is

Rate (°C/min) Temperature (°C) Hold time (min) 65.0 1.00 15.00 170.0 0.00 8.00 300.0 0.00 10.00 390.0 X

Table 4: Variation possibilities of the oven temperature program for optimization of retention times

21 III. PART 4 Biodiesel quality control according to DIN EN 14105

Signal identification

uV (x 10,000) Used naming Alternative naming 5.0 Glycerin “Free” Glycerin 4.5 Butanetriol (ISTD1) 4.0 Monopalmitin 1-Monopalmitoyl-rac-glycerol Tricaprin ISTD 2 Tricaprin 3.5 Monolinolenin 1-Monolinoleoyl-rac-glycerol Monoolein 1-Monooleoyl-rac-glycerol

3.0 -9-octadecenoyl]glycerol Butanetriol ISTD 1 1-Monooleoyl-rac-glycerol cis Monostearin 1-Monostearoyl-rac-glycerol 2.5 Glycerin

1,3-Di[ Tricaprin (ISTD2) 1,2,3-Tridecanolyglycerol 2.0 -9-octadecenoyl]glycerol cis Diglyceride 1,3-Di[cis-9-octadecenoyl]glycerol 1.5 Diglyceride (rest) All other diglycerides 1,2,3-Tri[ 1.0 Triglyceride 1,2,3-Tri[cis-9-octadecenoyl]glycerol 0.5 Triglyceride (rest) All other triglycerides

0.0 Table 1: Naming of biodiesel components -0.5 5.0 10.0 15.0 20.0 25.0 30.0

Minutes ince the standards contain Glycerin and are only one component of evaluated as single peaks and Figure 1: Standard chromatogram measured on a 25 m HT5, ID 0.32 mm, S each group (mono-, di- and grouped by the software. Di- and df = 0.1 µm with retention gap triglycerides, Figure 1) the identi- triglycerides are integrated as one fication of all components must group respectively. be carried out in a real biodiesel sample (Table 1). Identification

Four groups are evaluated in the The biodiesel chromatogram biodiesel quality control – “free” (Figure 3) starts with glycerin and glycerin, the monoglycerides, the first internal standard butane- diglycerides and triglycerides. triol. After the FAME peaks

uV (x 10,000) Four groups to be evaluated in biodiesel quality control 3.0 Glycerin Correct identification of monoglycerides is challenging 2.5

2.0

1.5

1.0

0.5 Butanetriol ISTD 1

3.25 3.50 3.75 4.00 4.25 4.50

Minutes

Figure 2: Glycerin and butanetriol peak zoomed from the biodiesel chromatogram in Figure 3

22 Biodiesel quality control according to DIN EN 14105 III. PART 4

(from 10 to 14.5 minutes) the uV (x 10,000) monoglycerides elute followed by

the second internal standard tri- Column Temp. (Monitor) caprin, and the di- and triglyc- 1.50 350 erides. 1.25 300 Monoolein + Monolinolein From the standard chromatogram 1.00 (Figure 1) the retention times of 250 °C glycerin, butanetriol (Figure 2) as 0.75 well as tricaprin are known. 200

0.50 Diglycerides 150

Additionally one component of Glycerin 0.25 each group Monostearin Monopalmitin Butanetriol ISTD 1 • 1-monooleoyl-rac-glycerol Triglycerides 100 (monoglycerides) 0.00 • 1,3-di[cis-9-octadecenoyl] 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 25.0 27.5 30.0 glycerol (diglycerides) Minutes • 1,2,3-tri[cis-9-octadecenoyl] glycerol (triglycerides) is identified in the standard chro- matogram. Figure 3: Biodiesel chromatogram measured on a SGE HT5, 25 m, ID 0.32 mm, df 0.1 µm with a retention gap 2 m, ID 0.53 mm In principle, four monoglycerides should be found in the chro- matogram. In case of the diglyc- uV (x 100,000) erides up to 10 separate peaks are in the chromatogram but for the 2.00 triglycerides seldom more than 1.75 two peaks are found. In the biodiesel sample (Figure 3) only 1.50

one major triglyceride peak was Monoolein (18:1,2,3) found. 1.25

1.00 Figure 4a shows the monoglyc- Monoolein + Monolinolein eride pattern stated in DIN EN 0.75 14105. For comparison Figure 4b 0.50

shows the corresponding part Monolinolenin (18:1,2,3) Monopalmitin (16:0) Monostearin

zoomed out of Figure 3. Accord- 0.25 Monopalmitin ing to DIN EN 14105 the peak 0.00 marked with a green arrow is 14.0 14.5 15.0 15.5 16.0 16.5 monolinolenin. In order to get a positive identification for all Minutes monoglycerides a mixture from Supelco was injected containing all four monoglycerides. The Figure 4a: pattern Figure 4b: Monoglycerides zoomed from the biodiesel chromatogram resulting chromatogram shows taken from DIN EN 14105 in Figure 3 three peaks only. Using the pure components a co-elution of monolinolenin and monoolein on the SGE HT5 column could be proven. ‘

23 III. PART 4 Biodiesel quality control according to DIN EN 14105

uV (x 1,000)

4.5

4.0

3.5 C-36:DG

3.0 -9-octadecenoyl]glycerol cis C54:TG

2.5 1,3-Di[

2.0

1.5 C52:TG

1.0 C-34:DG

0.5 C56:TG C58:TG C-38:DG 22.25 22.50 22.75 23.00 23.25 23.50 23.75

Minutes

Figure 5a: Diglyceride pattern taken Figure 5b: Diglycerides zoomed from the biodiesel chromatogram Figure 6a: Diglyceride pattern taken from DIN EN 14105 in Figure 3 from DIN EN 14105

The same result was found on four peaks can be found as shown Appendix A When ordering please specify custom- other non-polar columns (e.g. in the pattern given in DIN EN The mixture of monoglycerides from Sigma designed according to Lot No. DE1675 made by Zebron ZB 5HT). Therefore, 14105 (Figure 6a). Aldrich (Supelco) is custom designed. Sigma Aldrich Europe. a positive identification is recom- mended by using a monoglyc- Often only the signal of 1,2,3- It contains: The components in the ampoules can be eride mixture – for more infor- tri[cis-9-octadecenoyl]glycerol is • 1-Monopalmitoyl-rac-glycerol 25 % w/w dissolved in heptane. It may be treated as a mation see Appendix A. found. This makes the definition • 1-Monolinoleoyl-rac-glycerol 25 % w/w biodiesel sample. Just add internal standard of the triglyceride group difficult. • 1-Monooleoyl-rac-glycerol 25 % w/w and MSTFA for derivatization. The diglycerides appear as one • 1-Monostearoyl-rac-glycerol 25 % w/w group of peaks in the chromato- According to the DIN EN stan- • Total of 100 mg per ampoule gram (Figure 5a). In most cases dard the middle of the group the biggest peak is the 1,3-di[cis- corresponds approx. with the uV (x 1,000) 9-octadecenoyl]glycerol. Using retention time of the 1,2,3-tri- 10.5 this as marker the start and end [cis-9-octadecenoyl] glycerol peak of the “diglyceride peak group” plus 0.4 minutes. We recommend 10.0 can be identified according to the evaluating all peaks within the pattern shown in the DIN EN range of the resulting retention 9.5 standard (Figures 5a and 5b). time ± 1.5 minutes as triglycer- ides (Figure 6b). Dependent on 9.0 If identification of start and end column and temperature program -9-octadecenoyl]glycerol of the group is uncertain it is used also ± 1.0 minute could be cis better to implement one peak too sufficient. 8.5 many than one too less. Mostly 1,2,3-Tri[ the peaks are very small and do In Figure 6b the retention time 8.0 not contribute much to the final of 1,2,3-tri[cis-9-octadecenoyl] concentration. glycerol is 30.27 minutes, plus 7.5 0.4 yields 30.67 minutes. Within 28.5 29.0 29.5 30.0 30.5 31.0 31.5 32.0 It may happen that no triglyc- the interval from 29.17 to 32.17 eride signals can be found in a minutes all peaks were evaluated Minutes biodiesel chromatogram. Rarely as triglycerides.

Figure 6b: Triglycerides zoomed from the biodiesel chromatogram in Figure 3

24 Biodiesel quality control according to DIN EN 14105 III. PART 5

Evaluation and quantification

uV (x 10,000) Used naming Alternative naming Glycerin “Free” Glycerin 4.5 Monoolein 1-Monooleoyl-rac-glycerol

4.0 Diglyceride 1,3-Di[cis-9-octadecenoyl]glycerol Triglyceride 1,2,3-Tri[cis-9-octadecenoyl]glycerol 3.5 Table 1: Biodiesel components used to determine response factors 3.0 of the different groups -9-octadecenoyl]glycerol

2.5 cis 1,3-Di[ 2.0 ual time interval, retention time Settings for di- and triglycerides 1.5 ± ∆t, can be selected. Depending on the column used a “Default 1.0 Band Time” between 0.05 and 0.1 In Part 4 (Page 22 ff) of this man- min is sufficient. Relative reten- ual the identification of the di- 0.5 22.25 22.50 22.75 23.0 23.25 23.50 23.75 tion times should be selected as and triglycerides was discussed. “Identification Method” – espe- For quantification of these “peak cially if the samples are injected groups” GCsolution software 22.3 min 23.6 min manually. offers a special tool using the identification by band (Figure 1). Under “Peak” select “All Peaks” ‘ Figure 1: Diglycerides in a fter identification of the and for grouping “Concentration biodiesel chromatogram components in a biodiesel Summation” is selected. In the Achromatogram (Figure 3, next step the compound table has Part 4 on Page 23 of this applica- to be created – see Tab. 2 on Page tion manual) quantification will 26 as an example. The concentra- be the next step. According to tions of the calibrated compo- DIN EN 14105 a calibration nents are set according to Part 1 curve is measured for one com- (Page 15 ff) of this manual (sam- GCsolution “Band Method” pound of each group. The calcu- ple preparation). Uncalibrated for easy quantification of di- lated response factors are adapted components are set to 1. and triglycerides for all components of the respec- tive group. Under “Type” the Internal Stan- GCsolution “Grouping Function- dard components are marked as ality” calculates the required Representative for all other com- “Internal Standard and Refer- total glyceride content ponents, the compounds in Table ence” if relative retention times 1 are calibrated by determination are used. According to DIN EN Easy handling of multiple of a four level calibration curve. 14105 butanetriol is used as inter- internal standards nal standard for glycerin and tri- Quantitation parameters caprin for all other groups. This in GCsolution can be set under “Internal Stan- dard Groups” in the compound The application uses an Internal table (ISTD groups, Table 2). Standard calibration with four calibration levels (concentra- It is recommended to add the tions). The curve fit type is set same internal standard amount to to “linear”. The calibration curve all calibration standards as well as is not forced through zero. A to all unknown biodiesel samples. weighting method is not used and This would allow setting the the unit is mass-%. internal standard concentrations in the compound table equal to For identification the “Band” the “Internal Standard Amount” method is used i.e. for every of the unknown samples – the identified component an individ- easiest way is to set all to 1.

25 III. PART 5 Biodiesel quality control according to DIN EN 14105

Name Type Ret. time Band ISTD group Group ID No. Conc. 1 Conc. 2 Conc. 2 Conc. 2 Glycerin Target 3.442 Default 1 1 0.005 0.02 0.035 0.05 Butanetriol ISTD1 ISTD&Ref. 4.281 Default 1 0 1 1 1 1 Monopalmitin Target 14.642 Default 2 2 1 1 1 1 Monolinolein Target 15.921 Default 2 2 1 1 1 1 Monoolein Target 16.102 Default 2 2 0.2 0.8 1.4 2 Monostearin Target 16.35 0.02 2 2 1 1 1 1 Tricaprin ISTD2 ISTD&Ref. 20.488 Default 2 0 1 1 1 1 Diglyceride Target 23.154 Default 2 3 0.05 0.2 0.35 0.5 Diglycerides (rest) Target 22.95 0.65 2 3 1 1 1 1 Triglyceride Target 30.3 Default 2 4 0.05 0.2 0.35 0.5 Triglycerides (rest) Target 30.67 1.5 2 4 1 1 1 1

Table 2: Example of a compound table layout in GCsolution software. Clicking with the right mouse button on the compound table shows the function “Table Style.” This allows hiding of unnecessary columns and defini- tion of a convenient format.

Conc. ratio (x 100) According to the diglyceride pat- identified diglyceride it is recom- tern in DIN EN 14105 all peaks mended to quantify it separately. 4 0.020 in the retention time range from Therefore, the diglycerides have 22.3 to 23.6 minutes belong to two entries in the compound 0.018 the diglyceride group. The 1,3- table: diglyceride (1,3-di[cis-9- di[cis-9-octadecenoyl]glycerol is octadecenoyl]glycerol and di- 0.015 3 positively identified by calibra- glycerides (rest) representing all tion. The set retention time (RT) other components of this group. 0.013 in the compound table is the middle of this interval: It is important that diglyceride 0.010 ranges in the compound table 2 before diglycerides (rest). GCso- 0.007 RT + RT 22.3 + 23.6 RT = 1 2 = = 22.95 min 22 lution will then quantify diglyc- eride first and does not imple- 0.005 ment this peak in the diglycerides The band (∆t) is the difference (rest) group. 0.003 1 divided by two:

0.000 The calculation of the triglyc- 0.0 1.0 2.0 3.0 erides is performed in a similar RT - RT 23.6 - 22.3 Area ratio ∆ t =2 1 = = 0.65 min way. The determined retention 2 2 time interval from 29.17 to 32.67 minutes yields a retention time Y = aX + b Internal standard Since 1,3-di[cis-9-octadecenoyl] RT = 30.67 minutes and band a = 0.5415221 Calibration curve: Linear glycerol is the only positively ∆t = 1.5 minutes (Table 2). b = 4.398673e-003 Original: Not forced R^2 = 0.9998509 R = 0.9999255 Name RF Y(0) Glycerin 0.09161 0 Figure 2: Calibration curve for monoolein Butanetriol ISTD1 0 0 Monopalmitin 0.54152 0 Monolinolein 0.54152 0 Monoolein 0.54152 0 Monostearin 0.54152 0 Tricaprin ISTD2 0 0 Diglyceride 0.62761 0 Diglyceride (rest) 0.62761 0 Triglyceride 0.8165 0 Triglyceride (rest) 0.8165 0

Table 3: Response factor settings in the compound table after switching to Curve Fit Type: Linear RF (Linear)

26 Biodiesel quality control according to DIN EN 14105 III. PART 5

Group ID No. Name Unit Group No. Name Conc. Units 1 Glycerin mass % 1 Glycerin 0.0549 mass % 2 Monoglycerides mass % 2 Monoglycerides 0.3599 mass % 3 Diglycerides mass % 3 Diglycerides 0.0859 mass % 4 Triglycerides mass % 4 Triglycerides 0.0018 mass %

Table 4: Group table definition in GCsolution. Like all other tables in Table 6: Group results table in GCsolution software GCsolution the layout of the group table can be changed using the “Table Style” function available by right mouseclick on the respective table.

Name Conc. Units Peak No. triglyceride (rest) retention time With the results from Table 6 a Glycerin 0.0549 mass % 1 interval. total glyceride value of 0.159 is Butanetriol ISTD1 0 mass % 3 calculated. Thus, the content of Monopalmitin 0.024 mass % 7 A nice control option offers the free glycerin in the biodiesel sam- Monolinolein 0.0122 mass % 15 Peak No. row in this table. It ple measured is too high but all Monoolein 0.3007 mass % 16 shows that the peaks with the other requirements are met Monostearin 0.0229 mass % 17 numbers 33 to 40 are added including the total glyceride limit Tricaprin ISTD2 0 mass % 22 under diglyceride (rest) with the (Table 7). Diglyceride 0.0301 mass % 38 exception of peak number 38 Diglyceride (rest) 0.0558 mass % 33, 34, 35, 36, 37, 39, 40 which was already quantified as Summary Triglyceride 0.0018 mass % 45 diglyceride. In routine analyses the “step by Table 5: Compound results table in GCsolution software The group results table (Table 6) step” procedure described is, of finally contains the information course, automated in GCsolution Calibration tively (Table 3). All offsets “Y(0)” required for total glycerol calcu- software. The batch processing are set to zero. lation according to DIN EN functionality is a powerful tool in After measuring the four stan- 14105. the development of calibration dard samples with this method To calculate the total glyceride curves or quantification of a GCsolution calculates a calibra- value according to DIN EN The total glyceride value is calcu- sequence of unknown samples. tion curve for each group (Figure 14105 the sum concentrations for lated according to the equation: Using export functionalities the 2). It is important to check the all groups are needed. These are calculation of the total glyceride offset “b” of all calibration obtainable using another group- 1.000* Free glycerol value can also be automated. curves. If the deviation from zero ing functionality of GCsolution + 0.255* Monoglycerides is within acceptable tolerances software. In the compound table + 0.146* Diglycerides and the curve fit is also good, the (Table 2) the groups are defined + 0.103* Triglycerides response factors can be adapted under “Group ID No.” The = Total glyceride for the other group constituents. groups must then be named in the group table (Table 4). The Quantitation method method is now completed and Controlled contents Dimension Max. value Determination can be used for the evaluation of in biodiesel according to In the GCsolution “Calibration biodiesel samples. Monoglycerides % (m/m) 0.8 DIN EN 14105 Curve Window” the calculated Diglycerides % (m/m) 0.2 DIN EN 14105 response factors can be finally Results Triglycerides % (m/m 0.2 DIN EN 14105 applied to the respective compo- Free glycerin % (m/m) 0.02 DIN EN 14105 nents. Under “Quantitative” the After evaluation of a biodiesel Total glyceride % (m/m) 0.25 DIN EN 14105 “Curve Fit Type” is changed to sample GCsolution provides “Manual RF (Linear).” The cali- three different result tables. The Table 7: Limits for the content of the different glycerides and free glycerin bration curves disappear and in peak table lists all detected peaks according to DIN EN 14214 the compound table the “Conc.” – identified and not identified. rows are replaced by the rows The compound results table pro- “RF” and “Y(0).” RF is the res- vides concentrations for all com- ponse factor and equal to “a” in ponents listed in the compound Figure 2. Y(0) is the offset and table. equal to “b.” For glycerin, mono- olein, di- and triglyceride the RF Table 5 shows results from a values are set equivalent to the biodiesel sample containing values determined by the calibra- almost no triglycerides. Only one tion curves. These can be copied peak was detected for triglyceride to the other components respec- but no peak was detected in the

27 2.0/E/08/12/L me-dus.com

Shimadzu Europa GmbH Albert-Hahn-Str. 6-10 · D-47269 Duisburg

Tel.: +49 - (0)203 - 76 87-0 Fax: +49 - (0)203 - 76 66 25 Email: [email protected] To find your local Shimadzu contact please visit www.shimadzu.eu