Characterization of natural dyes by comprehensive 2D liquid chromatography

(1)

Bachelor Thesis Scheikunde

door

12 juli 2016

Studentnummer

10017089

Onderzoeksinstituut Verantwoordelijk docent

Van ‘t Hoff Institute for Molecular Sciences Dhr. Prof. dr. ir. P.J. Schoenmakers

Onderzoeksgroep Begeleider

(2)

Abstract

Characterization of dyes and their degradation products is essential for restoration and conservation of historical objects like paintings and textiles. Previously a method was developed to characterize synthetic dyes using a comprehensive 2D-LC method with ion-exchange chromatography in the first dimension and ion-pair reversed-phase chromatography in the second dimension. Synthetic dyes often contain negative charges, for this reason strong anion-exchange chromatography is used. This method was adjusted so that synthetic and natural dyes can be characterized simultaneously. Natural dyes are mainly neutral compounds causing them to elute quickly from the ion-exchange column. Stimulating the reversed-phase character of the ion-exchange column increases the retention time. By decreasing the organic modifier in the mobile phase, the hydrophobic retention is increased. The ratios water to acetonitrile 50:50, 60:40 and 70:30 [v/v] were tested; for the natural dyes the ratio 70:30 provided a better separation of the peaks over the chromatogram. However, applying this ratio on the synthetic dyes, not all the components elutes within the timeframe of the measurement. The dyes containing multiple negative charges were too strongly retained. By adjusting the ratio water to acetonitrile from 50:50 to 60:40 [v/v] in the first dimension, natural and synthetic dyes could be characterized simultaneously successfully.

Additionally a decrease in breakthrough peaks was achieved by decreasing the concentration of the organic modifier in the first dimension.

Samenvatting

Het kunnen analyseren van verfstoffen is van belang om informatie over historische objecten te verkrijgen zoals schilderijen en textiel. Kleurstoffen degraderen na verloop van tijd waardoor de objecten niet meer hun oorspronkelijke kleur hebben. Maar met informatie over de kleurstoffen en hun degradatie producten kan de originele staat van het object hersteld worden. In eerder onderzoek is een analyse methode ontwikkeld om synthetische kleurstoffen van elkaar te kunnen scheiden door middel van comprehensive twee dimensionale vloeistofchromatografie (2D-LC). Vloeistofchromatografie is een analyse methode waar stoffen van elkaar gescheiden worden op basis van verschillende eigenschappen. Door een kolom waar zich een stationaire fase bevind stroomt een mobiele fase heen en door interactie tussen de stof en de stationaire fase komen stoffen met verschillende eigenschappen op een ander moment de kolom uit. Deze verschillende tijden worden geregistreerd door een detector en in een chromatogram weergegeven als pieken. De detector bij deze methode gebruikt is een diode array detector, hiermee wordt van elke stof het absorptiespectrum gemeten. Bij kleurstoffen is dit een handig detectiemiddel, omdat elke kleur een ander absorptiespectrum heeft en ze hierdoor goed van elkaar te onderscheiden zijn.

Bij de methode om synthetische kleurstoffen te scheiden is anion-exchange en reversed-phase chromatografie gebruikt. Bij anion-exchange chromatografie worden deeltjes met meer negatieve ladingen langer vastgehouden op een kolom die positieve deeltjes op zijn stationaire fase heeft. Synthetische kleurstoffen bevatten vaak negatieve ladingen en daarom is voor deze scheidingstechniek gekozen. Bij reversed-phase chromatografie worden deeltjes op hun hydrofobe karakter gescheiden, hier blijven apolaire deeltjes langer in interactie met de apolaire stationaire fase van de kolom. Deeltjes die zich graag in water bevinden zullen hier dus weinig interactie aangaan

(3)

met de stationaire fase en snel van de kolom komen. Bij comprehensive twee dimensionale vloeistofchromatografie worden deze twee scheidingstechnieken achter elkaar toegepast; eerst gaan de deeltjes door de ion-exchange kolom (de eerste dimensie) en vervolgens door de reversed-phase kolom (tweede dimensie).

Tijdens dit project moest deze methode zo aangepast worden dat synthetische en natuurlijke kleurstoffen tegelijkertijd van elkaar gescheiden kunnen worden. Dit is nuttig wanneer er weinig kennis van een object is en men dus niet weet wat voor soort kleurstoffen gebruikt zijn. Door een universele analyse methode te hebben, hoeft deze voorkennis er niet te zijn.

Natuurlijke kleurstoffen hebben meestal geen ladingen en hierdoor zullen ze geen interactie met de stationaire fase van de ion-exchange kolom hebben. De anion-exchange kolom heeft een stationaire fase die bestaat uit positief geladen quaternaire ammonium deeltjes gebonden aan polymeren. Deze polymeren kan de stationaire fase ook interacties aangaan met hydrofobe deeltjes en heeft het dus een reversed-phase karakter. In de methode voor synthetische kleurstoffen was de mobiele fase zo gemaakt dat dit reversed-phase karakter geminimaliseerd was en de kleurstoffen voornamelijk op hun lading werden gescheiden.

Door de mobiele fase in de eerste dimensie aan te passen en hier een lagere concentratie organische vloeistof in te gebruiken, wordt het reversed-phase karakter van de ion-exchange kolom gestimuleerd en worden de deeltjes in de eerste dimensie niet alleen op lading gescheiden, maar ook op hun hydrofobe karakter. Hierdoor komen de neutrale kleurstoffen minder snel van de ion-exchange kolom af en ontstaat er betere scheiding van de kleurstoffen. Hieronder een afbeelding van de aangepaste methode waar natuurlijke en synthetische kleurstoffen van elkaar gescheiden kunnen worden. 0 25 50 75 100 125 150 175 200 225 250 275 300 113 100 88 75 63 50 38 25 13 0

First Dimension retention (min) - Strong Anion-Exchange

S e c o n d D im e n si o n r e te n tio n ( s) I o n -P a ir R e v e rs e d -P h a se C 1 8

Chromatogram van natuurlijke en synthetische kleurstoffen gescheiden door ion-exchange en reversed-phase chromatografie

(4)

1. Introduction

1.1 Dyes

Dyes are widely applied in many fields in industry such as in art, food coloration, cosmetics and histological staining.1_{Initially, dyes were obtained from different natural sources like plants, insects} and minerals. Not surprisingly, these dyes are thus commonly referred to as natural dyes. To achieve different colours, extractions from different plants are used. For example, madder, wild madder, bedstraws and safflower for red; weld, dyer’s broom, buckthorn species, sawwort and saffron for yellow; young fustic and sumac for yellow-brown; woad for blue; and alder and oak bark for black, greys or browns.2

Natural dyes are often hydrophobic particles that are neutral or slightly charged. Most natural dyes can be categorized into three dyestuff classes based on the dyeing techniques; mordant dyes, vat dyes and direct dyes.3_{Most of them are mordant dyes and the fibers have to be treated} with a metal salt to form a bond with the dyes. The mordant is a coordination metal that forms a complex with the fiber and the dye. The mordant forms bonds with neighboring carbonyl and/or alcohol groups in the dye. Iron and aluminum were often used as mordant, later also tin, chromium and copper. The chosen mordant can also have an influence on the colour obtained.2_{Direct dyes} form a bond with a fiber without any pre-treatment through hydrophobic interactions, hydrogen bonds and ionic interactions.2_{Vat dyes are not soluble in water and become water-soluble by} reduction. This reduction was traditionally achieved by fermentation, but nowadays this is achieved using reducing agents like sodium dithionite.2

In 1856 William Henry Perkin discovered mauveine, the first synthetic dye. The discovery of synthetic dyes lead to the rapid replacement of the natural dyes by synthetic dyes between 1870 and 1880.3_{Synthetic dyes belong to the acidic, basic direct and mordant dyestuff classes. Acidic and basic} dyes are water-soluble anionic and cationic dyes.4

It is essential to be able to analyse dyes in historical objects, like paintings and textiles, to obtain information about the period they were made and their original condition. Information about dyes and their degradation products is important for the conservation of these degraded historical objects. An analytical method used for colour identification is high-performance liquid chromatography (HPLC).3

1.2 Liquid Chromatography

Liquid chromatography (LC) is an analytical separation technique consisting of a liquid mobile phase and a solid stationary phase in a column. The stationary phase remains in place inside the column while the mobile phase moves through the column. The partitioning of solutes between the mobile and stationary phase will lead to separation of the different components.5_{The interaction of the} solutes with the stationary phase depends on the properties of the solutes and of the column used. The liquid entering a column is called eluent and liquid leaving a column is called eluate.5

Using detectors the solutes leaving the column can be observed in a chromatogram. The peaks in the chromatogram represent the retention times and the detector response of the solutes,

(6)

for example UV-absorbance. The retention time (tr) is the time needed for the component to reach the detector after being injected. The time needed by an unretained component or the mobile phase to react the detector is called tm. With these two factors the retention factor (k) can be calculated using the formula:

k =

t

r

−

t

m

t

m

.5_{This retention factor shows the ratio of time the solute spends in} the stationary phase and time spends in mobile phase, so components with a large retention factor are stronger retained and elute slower from the column than components with a small retention factor.

Peak capacity is the number of peaks that can be separated within a retention window.6_This number depends on the peak widths, and more complex mixtures can be separated if the peak capacity is high. Peak broadening is a factor that decreases the peak capacity and leads to a poor separation. The plate height (H) gives information about the peak width; a smaller plate height equals a narrower peak.

The van Deemter equation shows the relation between plate height and the linear flow rate (ux) and column used:

H ≈ A+

B

u

_x

+

C u

x .5 A is the multiple path term, which is independent of

the flow rate. Peaks broaden because particles of the same compound do not follow the same path through the column, which makes them elute at different times. B is the longitudinal diffusion, when a higher flow rate is used, the zone of solute elutes faster from the column and has less time to diffuse and broaden the peak. The C-term arises from the equilibration time of the solute between the stationary and mobile phase. When this equilibration is slow, a faster flow rate will broaden the peak.

The particle size of the stationary phase has influence on the plate height; small particles give better resolution. This is illustrated in Figure 1 by the van Deemter curves of several particle sizes.

Figure 1: Plate height as a funtion of flow rate for stationary phase particle diameters of 10, 5 and 3 μm (Source: Harris, D. C. Quantitative Chemical Analysis. (W.H. Freeman and Company, 2010), p. 556)

However when decreasing the particle size of the stationary phase a higher pressure is required to achieve the same flow rate. HPLC can operate at pressures up to 400 bar and to improve the plate height and the run time ultra performance liquid chromatography (UPLC) can be used; UPLC can operate at pressures up to 1200 bar.

(7)

1.2.1 Reversed-phase liquid chromatography

Reversed-phase liquid chromatography (RPLC) consists of a nonpolar stationary phase and a polar mobile phase and therefore a less polar solvent has higher eluent strength.5_{Components that have a} hydrophobic character are retained, while components with a hydrophilic character will elute quickly as they stay in the polar mobile phase and do not interact with the hydrophobic stationary phase.

1.2.2 Ion-pair RPLC

Ion-pair reversed-phase liquid chromatography (IP-RPLC) is used when the sample contains ionic components. As these components have charges they are hydrophilic and they can stay in the mobile phase. Then no separation can be made because the components would not be retained through interaction with the hydrophobic packing of the column. But with ion-pair RPLC a counterion is added to the mobile phase and this counterion shields the charge on the ionic components and creates a hydrophobic character for the components that leads to retention.

As dyes contain negative charges, the counterion is a cation that shields the negative charge of the dye by forming an ion pair. Tetramethylammonium (TMA), tetrabutylammonium (TBA) and sodiumhydroxide (NaOH) were tested and TMA was better at dispersing the peaks in the

chromatogram, Figure 4.4_{The use of TMA in the mobile phase compared to NaOH led to better}

spreading of the peaks over the chromatogram and a higher resolution. The use of TBA did not improve the separation obtained when using TMA, the spreading between peaks decreased. In the reversed-phase gradient the amount TMA decreases and the concentration acetonitrile increases, this will transfer charged components into the mobile phase.

Figure 2: Chromatograms of the separation of synthetic dyes using TBA (A), TMA (B) and NaOH (C)) in the mobile phase.

(Source: Pirok, B.W.J., Knip, J., van Bommel, M. R. & Schoenmakers, P.J. Characterizaion of synthetic dyes by comprehensive two-dimensional liquid chromatography combining ion-exchange chromatography and fas ion-pair reversed-phase chromatography, J. Chromatogr. A 1436, 144 (2016).)

1.2.3 Ion-exchange chromatography

Ion-exchange chromatography (IEX) separates components based on their charge. The strength of the interaction between the stationary phase and the component depends on the number of the charges on the component. IEX has four subdirectories: weak anionic exchange, strong anionic

(8)

exchange, weak cationic exchange and strong cationic exchange. In anionic exchange the IEX column consists of cationic packing that can retain negatively charged analytes. Strong anionic exchange (SAX) columns contain -NR3+ groups. Weak anionic exchange columns contain -NH2, -NHR or -NR2 groups.7 _{Cationic exchange has columns with negatively charged packing to retain positively charges} analytes. Strong cationic exchanger contains the -SO3- ionogenic group and weak cationic exchange columns contain -CO2H, -PO3H2, -OH, or -SH.

By increasing the salt concentration in the mobile phase, first the components with the least amount of charges will elute from the column, because these components have a weaker interaction with the stationary phase. The strongly interacting components that contain the most charges need a higher concentration of salt to be eluted from the column. Using a gradient during the run that increases the salt concentration would lead to an elution of the least charged components first and the most charged components last. Adjusting the pH to elute components from the column is only applicable for weak ion-exchange columns, the pH will neutralize the column or the components and the charged components will elute. Synthetic dyes are often negatively charged; therefore a SAX column is used in the first dimension.

1.3 LCLC

There are two types of two-dimensional liquid chromatography: heart-cutting 2D-LC (or LC-LC) and comprehensive 2D-LC (or LC×LC). Only a selected fraction of the eluate from the first dimension enters the second dimension when using LC-LC, see Figure 2. Meanwhile, using LC×LC all the eluate of the first dimension in the second dimension.

Figure 3: Comprehensive and heart-cutting 2D-LC

(Source: Pat Sandra, Gerd Vanhoenacker, ‘Demystifying Two-Dimensional Lquid Chromatography: Moving from 1D to 2D liquid chromatography is a big step towards the high peak capacities demanded by complex sample analysis’, the Analytical Scientist, issue 0314)

(9)

Passing all the eluate from the first dimension through the second is possible when the first dimension uses a low flow rate and the second dimension a high flow rate. The two dimensions are connected through a valve, Figure 3. The eluate from the first dimension fills the loop on position 1 and 8. When this loop is filled the valve switches and the content of the loop on position 1 and 8 is transferred to the column of the second dimension. In the mean time the eluate from the first dimension fills the loop on position 4 and 5. The filling and the flushing of the loops have to run simultaneously so nothing of the eluate of the first dimension will be wasted and everything is separated in the first and second dimension. To make simultaneous running possible the flow rate of the first dimension has to be low enough for the second dimension to pass through the column and detector. This way all the fractions follow each other continuously and a two-dimensional chromatogram can be obtained.

FIgure 3: Flow paths through valve for LC×LC

(Source; Peter W. Carr and Dwight R. Stoll, Two-dimensional liquid chromatography; principles, practical implementation and applications, primer, Agilent Technologies, 2015, p. 77.)

The columns used in the dimensions have to separate components on different properties, when using the same separation technique twice, 2D-LC will not have an added value for the chromatogram, no extra or new information will appear. But when two types of separation techniques used are orthogonal, a better separation will be achieved. In this study, an ion-exchange column and a reversed-phase column are used.

1.4 Separation of dyes

Most dyes have structures based on anthraquinones, flavonoids, indigoids, neo-flavonoids and carotenoids, as seen in Figure 4. As the molecular skeletons of many dyes are often made up of one of these five structures, some dyes look very similar, leading to similar LC retention times. Beside the structure similarity, a major reason why one-dimensional LC is insufficient is the complexity of samples from historical object. There is an immense amount of dyes, each containing impurities and degradation products and a sample obtained from a historical object will contain approximately

(10)

one-dimensional LC and an in practice applicable retention window because the peak capacity is too small.8_{However, using comprehensive 2D-LC the peak capacity increases significantly and separation} of samples containing many components can be achieved.

In previous work a method was developed for the separation of synthetic dyes, where a mixture of 54 different dyes was used.4_{In the first dimension, a strong anionic exchange (SAX) was} applied since most synthetic dye components are negatively charged. In the second dimension, ion-pair reversed-phase chromatography was used to separate the dyes based on hydrophobicity.

Figure 4: Main structures dyes

In this study, the feasibility of analyzing the natural dyes with the same, or similar, method will be investigated. To accomplish this, the method must be adapted to characterize both synthetic and natural dyes simultaneously. One problem of the previously developed method is the first dimension ion-exchange mechanism. Natural dye components are often hydrophobic particles that are neutral or have a small amount of charges. In ion-exchange, components are separated based on their charge. Therefore, the neutral natural dyes would not be retained by the stationary phase, which results in immediate elution from the column. However, the previously developed method used in the first dimension a strong anionic column with stationary phase containing quaternary ammonium groups covalently linked to a chemically stable polymer. These polymers contributed to mixed-mode chromatography; the quaternary ammonium leads to ion-exchange interactions and the polymers lead to reversed-phase interactions. With this mixed-mode chromatography solutes are retained on their charge as well as their hydrophobicity. In the previous developed method, the mobile phases in the first dimension contained water to acetonitrile in a ratio 50:50 to minimize the hydrophobic interactions with the stationary phase.4

This reversed-phase character can be used to increase the area where the natural dyes will elute so a better spreading between the peaks of the natural dyes can be achieved. Decreasing the concentration of organic modifier in the mobile phase will stimulate the hydrophobic retention. In this project, the compatibility of the synthetic dyes method was investigated using 22 natural dye samples obtained from the reference collection of the National Heritage Agency. Then the organic modifier concentration of the mobile phase of the first dimension is decreased to see if this leads to higher peak resolution. In this study the results will be presented.

(11)

2. Experimental

2.1 Instrumental

The column used in the first dimension with anion-exchange was an Agilent PL-SAX column, PL1951-3802, 150 × 2.1 mm i.d., 8μm particles, 1000-Å pore size. An Agilent 1290 Inifinity in-line filter was placed before the anion-exchange column to prevent any solid particles entering the column. In the second dimension with reversed-phase an Agilent ZORBAX Eclipse Plus C18, Rapid Resolution HT 959941-902, 50 × 4.6 mm i.d., 1.8 μm particles was used.

For the start up measurements a Shimadzu HPLC system was used, from now on referred to as system 1. Both columns were tested to see if the same separation as during the method development could be achieved, so ion exchange and reversed-phase. The system consisted of a DGU-20A5 prominence degasser, FCV-10AL VP mixer, LC-10AD VP pomp, SPD-10A VP UV-VIS detector, CBM-20A Prominence communication bus module and the reservoir tray containing injection valve with cat. No. 228-34736-91. The system was controlled by a computer with Labsolution LCsolution Version 1.21 SP1 software.

Another HPLC system, from now on referred to as system 2, was used to measure the retention times of the separate dyes in reversed-phase and obtaining their absorption spectra. System 2 consisted of an Agilent 1100 Series G1311A Quatpump Serial # DE83105407, Agilent Technologies 1260 infinity G1322A 1260 degasser Serial No. JPAAJ82423, Agilent 1100 Series G1313A ALS Serial # DE23920534, HP Hewlett Packard series 1100 G1316A COLCOM serial # DE 91609277 and Agilent Technologies 1290 infinity G4212A 1290 DAD (diode-array detector) Serial No. DEBAF02289. The system was controlled by a computer with Agilent OpenLAB CDS Chamstation Edition (Rev. C.0104 [35]) software.

Agilent 1290 Infinity was used to take 1D ion exchange chromatograms and 2D LC chromatograms and consisted of G4220A 1290 Bin Pump Serial No. DEBAA03255 (first dimension), G4226A 1290 Sampler Serial No. DEBAP03712, G1316C 1290 TCC (thermostat column compartments) Serial No. DEBAC05647 (first dimension), G4220A 1290 Bin Pump Serial No. DEBAA03257 (second dimension), G1316C 1290 TCC Serial No. DEBAC05692 (second dimension) and G4212A 1290 DAD Serial No. DEBAF02307. The loops connected to the valve had a volume of 60 μL. The system was controlled by a computer with Agilent OpenLAB CDS Chemstation Edition (Rev. C.0104 [35]) software.

2.2 Chemicals

Only deionized water was used during the experiments (Arium 611UV, Sartorius, R=18.2 MΩcm, Germany). Methanol absolute (ULC/MS - CC/SFC grade), 2-propanol (ULC/MS - CC/SFC grade) and Acetonitrile (AR grade, ULC/MS - CC/SFC grade and LC/MS grade) were obtained from Biosolve (Valkenswaard, the Netherlands). Tris(hydroxymethyl)aminomethane hydrochloride (99+%) was obtained from Acros Organics (New Jersey, USA / Geel, Belgium). Ammonium Sulfate (BioXtra, ≥ 99.0%), Tetramethylammonium hydroxide (25 wt. % in H2O), Formic acid (reagent grade, ≥95%), Trisma Base (tris[hydroxymethyl]aminomethane) (Reagent Grade; Minimum 99.9%) and Methylsulfoxid (>=99%) were obtained from Sigma-Aldrich (Steinheim, Germany).

(12)

Heritage Agency of the Netherlands; see Table 1 for an overview of the samples and see Appendix Table 2 and 3 for an overview of the 22 dye samples and their structures.

Table 1: Overview natural dyes

Sample number Dye

0228 Sandelwood 0670 Purpurin 0894 Kaempferol 0895 Emodin 1153 Brasilin 1154 Maclurin 1156 Turmeric 1164 Berberine 1166 Fisetin 1167 Rhamnetin 1168 Carminic acid 2029 Lawson 2031 Indigo 2032 Morin 2035 Quercetin 2036 Rutin 2961 Orcein 2989 Isatin 3177 Haematein 4796 (+)epicatechin 4799 (+)catehinhydrate 7200 Campechewood

2.3 Analytical conditions

2.3.1 Sample preparation

To keep the method compatible with the one previous developed, the dye samples were dissolved to form a 100 ppm solution per dyestuff when combined. To achieve this concentration the separate samples where dissolved in water to methanol 1:1 [v/v] to form a 2000 ppm solution, so when all 22 samples were combined a mixture of approximately 100 ppm was formed. 1000 ppm solutions were made from samples 0895 and 2031, because there was not enough of the samples available to make a higher concentration and obtain a solution with a minimal volume of 1 mL.

For the measurements in 2D-LC a mixture of synthetic dyes, a mixture of natural dyes and a 1:1 mixture of the synthetic and natural mixtures were used. The natural dyes were measured separately in the 1D-LC measurements.

2.3.2 Methods

On system 1 both anion-exchange and reversed-phase separation were measured one-dimensional. For the anion exchange separation the Agilent PLSAX column was used with a flow rate of 0.5

(13)

mL/min. Mobile phase A consisted of water to acetonitrile 1:1 [v/v] and mobile phase B 100 mM ammonium sulfate in water to acetonitrile 1:1 [v/v]. Measurements with mobile phases with water to acetonitrile 60:40 [v/v] and 70:30 [v/v] were also performed. The time program used was 0.0-0.5 min isocratic at 100% A, 0.5-10.5 min linear gradient to 100% B, 10.5-14.5 min isocratic at 100% B, 14.5-15.0 min linear gradient to 100% A, 15.0-16.0 min isocratic at 100% A. UV light at 254 & 370 nm or 371 & 450 nm. When using the 70:30 mobile phase, the measuring time was increased by adding 15.0-45.0 min isocratic at 100% A to the timetable. The injection volume used was 20 μL. For the reversed-phase measurements the Agilent ZORBAX column was used with a flow rate of 0.75 ml/min. The flow rate used in the previous developed method is 1.0 mL/min, but the tubing on the Shimadzu system could not operate at pressures above 200 bar so the flow rate was adjusted. Mobile phase A consisted of a 10 mM TMA and 66 mM formic acid in water to acetonitrile 95:5 [v/v] and mobile phase B of a water to acetonitrile 5:95 [v/v]. The time program used was 0.0-0.5 min isocratic at 100% A, 0.5-12.5 min linear gradient to 100% B, 12.5-13.0 min isocratic at 100% B, 13-14.5 min linear gradient to 100% A, 13-14.5-15.0 min isocratic at 100% A. The injection volume used was 20 μL.

On system 2 one-dimensional measurements were obtained doing reversed-phase and anion exchange separation. For the anion exchange separation the Agilent PLSAX column was used with a flow rate of 0.5 mL/min was used. Mobile phase A consisted of water to acetonitrile 1:1 [v/v] and mobile phase B 100 mM ammonium sulfate in water to acetonitrile 1:1 [v/v]. Measurements were performed with water to acetonitrile 50:50 [v/v] and water to acetonitrile 90:10 [v/v]. The time program used was 0.0-0.5 minute isocratic at 100% A, 0.5-10.5 min linear gradient to 100% B, 10.5-14.5 min isocratic at 100% B, 10.5-14.5-15.0 min linear gradient to 100% A, 15.0-16.0 min isocratic at 100% A. When using mobile phases with water to acetonitrile 90:10 [v/v] the time program was extended to 45 minutes.

For the reversed-phase measurements the Agilent ZORBAX column was used with a flow rate of 1.0 ml/min. Mobile phase A consisted of a 10mM TMA and 66 mM Formic acid in water to acetonitrile 95:5 [v/v] and mobile phase B of water to acetonitrile 5:95 [v/v]. The time program used was 0.0-0.5 min isocratic at 100% A, 0.5-12.5 min linear gradient to 100% B, 12.5-13.0 min isocratic at 100% B, 13-14.5 min linear gradient to 100% A, 14.5-15.0 min isocratic at 100% A. Using the Agilent Infinity system for the LC×LC measurements the Agilent PLSAX column was used in the first dimension and mobile phase A consisted of water to acetonitrile 1:1 [v/v] buffered at pH=7.5 with 0.6572 g/L TRIS.HCL and 0.1006 g/L TRIS base. Mobile phase B consisted of 100 mM ammonium sulfate in water to acetonitrile 1:1 [v/v] buffered at pH=7.5 using 0.6572 g/L TRIS.HCL and 0.1006 g/L TRIS base. The flow rate was set at 0.01 mL/min and the time program used was 0.0-10.0 min isocratic at 100 % A, 0.0-10.0-190.0 min linear gradient to 100% B, 190.0-200.0 min linear gradient to 100% A.

In the second dimension the Agilent ZORBAX column was used with a flow rate of 2.4 ml/min. Mobile phase A consisted of a 10mM TMA and 66 mM Formic acid in water to acetonitrile 95:5 [v/v] and mobile phase B of a water to acetonitrile 5:95 [v/v]. The time program used was 0.0-1.5 min linear gradient from 100 % A to 100 % B, 0.0-1.5-1.6 min linear gradient to 100 % A and 1.6-2.0 min isocratic at 100 % A. The injection volume used was 20 μL.

(14)

3.1 IP-RPLC

3.1.1 IP-RPLC of synthetic dyes

The synthetic dyes were not measured separately during this project; only the mixture in which all 54 dyestuffs were combined was used. This step was performed in order to see if the system was functional and the results from the previous study could be repeated. Also, this was used to check column efficiency; the first column had poor resolution; the second column was damaged and could not operate at the desired flow rate, as the pressure increased sharply immediately; the third column achieved a separation with good resolution and compatible peaks, see Figure 5. Because of the good separation when using column 3, this column was used during the rest of the experiment for reversed-phase separation.

Object 7

Figure 5: Synthetic dyes mixture (ten times diluted) separation at 254 nm using column 1 and 3

3.1.2 IP-RPLC of natural dyes

One-dimensional chromatograms were obtained from the mixture of natural dyes, Figure 6, as well as the separate dye samples, see Appendix Figures 20 up to 65 for the chromatograms and the UV-spectra of the peaks. Every dye absorbs at, sometimes only slightly, different wavelength depanding on which colour it absorbs. By obtaining the UV-spectra of all the separate dyes, the peaks in a chromatogram of the mixture can be identified. There are not 22 different peaks visible in the chromatogram of natural dyes, which means some dyes overlap with one another. This is a reason to apply 2D-LC; the second dimension would separate the peaks that have overlap and provide a higher peak resolution.

The chromatograms of the separate dyes contain several peaks, which should not be the case when a pigment is pure. Figure 5 also displays a blank run where only acetonitrile was injected; here

(15)

several peaks were visible. This must be because of impurities in the acetonitrile that absorb UV. In almost all the chromatograms there were several peaks visible apart from these impurities.

Object 9

Figure 6: Reversed-phase separation of natural dyes and acetonitrile

Natural dyes samples are purified and concentrated, but often there are still traces from other pigments that originate from the same source, for example morin, maclurin, quercetin and kaempferol. All these pigments originate from fustin.3_{The UV-spectra of the peaks show the} presence of maclurin, morin and kaempferol in the sample of maclurin when a chromatogram is obtained, shown in Figures 7 and 8. The peaks that belong to maclurin and kaempferol are the most intense and therefore the dye contains mainly maclurin and kaempferol.

Object 11

(16)

Object 13

Figure 8: UV-spectra 1154 maclurin

Sometimes several peaks are obtained because of the presence of isomers in the dye, for example in indigo; the presence of indigotin and indirubin is detected, Figures 9 and 10. Indirubin is an isomer of indigotin. The indigo sample was dissolved in DMSO because of the low solubility in methanol and water. Chromatograms of indigo dissolved in water and methanol are also obtained, but the intensity of the peaks is very low. However, this solution is used in the mixture of natural dyes so intense peaks from indigo will not appear in the natural dyes chromatograms.

Object 15

(17)

Object 17

Figure 10: UV-spectra 2031 indigo

A large amount of peaks are visible in the orcein chromatogram, Figure 11, as orcein is a dye consisting of a mixture of various amino and hydroxy orcein structures and their isomers, see Appendix Table 3. The UV-spectra of these peaks are all similar to each other; they only differ in retention time.

(18)

Figure 11: Reversed-phase separation 2961 orcein

3.2 IEX

After obtaining all the information of the dye samples in one-dimensional reversed-phase separation, the next step was to optimize the same parameters for ion exchange separation. The anion-exchange measurements of the natural dyes showed chromatograms with a few peaks using the Shimadzu system. Different ratios water to acetonitrile in the mobile phases were tested, and as expected the peaks in the chromatograms shifted to the right and thus the components were stronger retained, see Figure 12. When changing to system 2, no useful chromatograms could be obtained. When switching to the Agilent Infinity system on one-dimension settings the peaks appeared to be spread out over the whole chromatogram when looking at the absorbance. Along the whole chromatogram absorbance was present. Because of the unclear chromatograms in IEC, the 2D-LC measurements were started. The dye samples overlap using 1D-LC and the 22 dyes cannot be distinguished in these chromatograms of the natural dyes, this is why 2D-LC has to be applied.

Object 23

(19)

3.3 Comprehensive 2D-LC

First the 2D-LC system was tested with a LC×LC run of the mixture of synthetic dyes to validate the system by comparing it by previous obtained chromatograms. However, no peaks of the dye components could be observed because of enormous absorption of mobile phase B, Figure 13. This problem was remedied by changing the grade of the acetonitrile. First an AR grade was used, which is 99-percentage pure acetonitrile, but now ULC/MS-CC/SFC grade acetonitrile is used. The one percentage of impurity in the acetonitrile AR grade appears to have strong absorption. The absorption in the chromatogram corresponds with the gradient of mobile phase B in the second dimension. After changing to ULC/MS-CC/SFC grade the LC×LC chromatogram could be obtained that corresponded with the chromatograms from previous research, Figure 14. Using the same previous developed method chromatograms of the natural dyes and the mixture of natural and synthetic dyes combined were obtained, Figures 15 and Figure 65 up to 71 in Appendix.

Object 25

Figure 13: Second dimension chromatogram of 2D-LC measurement of mixture of synthetic dyes using acetonitrile AR grade

(20)

0 25 50 75 100 125 150 175 200 225 250 275 300 113 100 88 75 63 50 38 25 13 0

(21)

In Figure 14 three groups of peaks can be distinguished, the group on the left side which elute first from the column in the first dimension contains the dyes with +1 or no charge. The area with peaks in the middle of the chromatogram contain charges of -1 and the group of peaks on the right which are most retained contain the dyes with a charge of -2. The SAX column retains the dyes with a larger negative charge stronger and these dyes elute slower from the columns. The 0/+1 charged dyes have tailing in the first dimension, while the -2 charged dyes experience tailing in the second dimension.

The previous developed method to separate synthetic dyes also appears to work for the natural dyes; the expectation was that the natural dyes would elute practically immediately form the first column. However, the separation between the peaks is already reasonable. Using the UV-spectra in the 2D-LC chromatogram and the previous obtained UV-UV-spectra most of the natural dyes could be identified, Figure 15.

The peak identified as morin is spread out over a large area of the second dimension; this line is not visible in all the chromatograms. The chromatograms where this line does not appear over the whole first dimension were obtained within a few runs after column storage. Before the columns were stored they were flushed with storage liquids that cleaned them completely. The component morin appears to be strongly retained by the column.

The peak of kaempferol consists of three successive peaks, however all the UV-spectra are the same and belong to kaempferol. A reason for the splitting within the peak of one component could be the presence of kaempferol in other dye samples, for example in maclurin. Apparently these structures of kaempferol differ slightly and are retained separately.

(22)

To obtain more retention for the natural dyes, the ratio of water to acetonitrile in the first dimension was adjusted by decreasing the organic modifier. Instead of water to acetonitrile 50:50 [v/v] a ratio of 60:40 [v/v] and 70:30 [v/v] was used, see Appendix Figures 66-71 for the chromatograms. The different ratios were tested on all three the dye mixes to see the effect on both the natural as the synthetic dyes.

Comparing the chromatograms of the different water to acetonitrile ratios, it appears that the natural dyes have a better separation at 60:40 [v/v] than at 50:50 [v/v]. The difference in separation between the different ratios is visualized in the overlays of the chromatograms; see Figures 16 and 17. The decrease of organic modifier causes a shift of the peaks; the solutes are longer retained. However, this varies per solute, some component do not change significantly in retention time, others double in retention time. All the components elute within the timeframe of the measurement, so this ratio can be used to increase the resolution of the peaks.

Figure 16: Overlay of LC×LC chromatograms mixture of natural dyes using different concentration organic modifier In the overlay of the chromatograms of the separation of the synthetic dyes, Figure 17, the increased retention time is observed. However, using the ratio of water to acetonitrile 70:30 [/v] the peak with two negative charges do not elute within the timeframe from this method. To detect these

components the measure time has to be significantly increased and as this run already takes 5 hours and increasing this will not be used in practice.

Adjusting the method to a ratio of water to acetonitrile 70:30 [v/v] is not an option as characterization method since some synthetic dyes are retained too strongly. As a result, the best method to characterize natural and synthetic dyes simultaneously is the ratio 60:40.

(23)

Figure 17: Overlay of LC×LC chromatograms mixture of synthetic dyes using different concentration organic modifier

In addition to the significant shift of the peaks using different water to acetonitrile ratios, a decrease in breakthrough peaks can be observed in the synthetic dyes chromatograms. These peaks elute at t0.

The breakthrough peaks shift because of the water to acetonitrile ratio, but the intensity of the visible peaks also decreases, Figure 21. The chromatogram of separation using water to acetonitrile 50:50 [v/v] contains distinctly more and intense peaks in the t0 line in the chromatogram compared

to the ratios 60:40 and 70:30.

Figure 18: Breakthrough peaks in synthetic dyes chromatograms

A breakthrough peak appears on a chromatogram because the solvent in which the sample is dissolved is stronger than the mobile phase.9_{Because of the strong solvent in which the sample is} located, a part of the sample will remain in this solvent without interacting with the stationary phase and elute from the column at or near t0 additional to the real peak.

(24)

concentration, a more polar mobile phase is obtained. The dissolved sample will diffuse more easily to the mobile phase with a higher water concentration and will be retained by interactions with the column. Reducing the presence and intensity of the breakthrough peaks contributes to a clearer chromatogram and a higher intensity of the actual peaks of the components.

(25)

4. Conclusion

After obtaining retention times and UV-spectra of natural dyes, the previous developed 2D-LC method was applied on natural dyes. To see if this separation could be improved the ratio of water to acetonitrile in the first dimension was adjusted to 60:40 [v/v] and 70:30 [v/v]. The decrease of the organic modifier caused longer retention times for most of the components. However, using the ratio 70:30 [v/v] the retention times of some of the synthetic dyes exceeded the duration of the measurement. For this reason, this ratio is not applicable as method for the characterization of synthetic and natural dyes. A method to characterize synthetic and natural dyes was successfully developed by adjusting the ratio 60:40 [v/v] in the mobile phases in the first dimension. This ratio leads to a greater separation between peaks within the timeframe of the measurement.

Additionally, it was found that decreasing the organic in the first dimension led to a decrease in breakthrough peaks. The presence and intensity of the breakthrough peaks decreased because of the more polar mobile phase.

In future research coupling of the 2D-LC with an MS can help confirm the by UV-spectra identified peaks. Also this can help to identify the unidentified peaks in the chromatograms and clarify degradation products. This can also be helpful to clarify for example the peak of kaempferol, which consists of three successive peaks. The effect of other solvents in the mobile phases could be a suggestion for future research. Also the concentration organic modifier can be varied to investigate

the breakthrough phenomenon further.

This study indicates that so far the best method to characterize both natural and synthetic dyes is comprehensive 2D-LC with ion-exchange chromatography in the first dimension using mobile phases with water to acetonitrile 60:40 [v/v] and reversed-phase chromatography in the second dimension.

(26)

References

1. Shahid, M., Shahid-Ul-Islam & Mohammad, F. Recent advancements in natural dye

applications: A review. J. Clean. Prod. 53, 310–331 (2013).

2. Kirby, J., Van Bommel, M. & Verhecken, A. Natural Colorants for Dyeing and Lake Pigments;

Practical Recipes and Their Historical Sources. (Archetype Publications, 2004).

3. Hofenk de Graaff, J. H. The Colourful Past; Origins Chemistry and Identification of Natural

Dyestuffs. (Abegg-stiftung and Archetype Publications Ltd., 2014).

4. Pirok, B. W. J., Knip, J., van Bommel, M. R. & Schoenmakers, P. J. Characterization of synthetic dyes by comprehensive two-dimensional liquid chromatography combining ion-exchange chromatography and fast ion-pair reversed-phase chromatography. J. Chromatogr. A 1436, 141–146 (2016).

5. Harris, D. C. Quantitative Chemical Analysis. (W.H. Freeman and Company, 2010).

6. Gu, H., Huang, Y. & Carr, P. W. Peak capacity optimization in comprehensive two dimensional liquid chromatography: A practical approach. J. Chromatogr. A 1218, 64–73 (2011).

7. Katz, E., Eksteen, R., Schoenmakers, P. & Miller, N. Handbook of HPLC. (Marcel Dekker, Inc., 1998).

8. van Bommel, M. R., Berghe, I. Vanden, Wallert, A. M., Boitelle, R. & Wouters, J. High-performance liquid chromatography and non-destructive three-dimensional fluorescence analysis of early synthetic dyes. J. Chromatogr. A 1157, 260–272 (2007).

(27)

Appendix

Table 2: Dye samples

Sample number Trivial name Scientific name of biological source Application method

0228 Sandelwood Pterocarpus santalinus L. Mordant dye

0670 Purpurin Rubia tinctorum L. Madder;Mordant dye

0894 Kaempferol Rhamnus infectoria L. Persian berries; Mordant dye

0895 Emodin From different species Alderbark; Mordant dye

1153 Brasilin Caesalpinia sappan L. Mordant dye

1154 Maclurin

Chlorophora tinctoria L. / Maclura

tinctoria L. / Morus tinctoria L. Fustic; Mordant dye

1156 Turmeric Reseda luteola L. Direct dye

1164 Berberine Berberis vulgaris L. Direct dye

1166 Fisetin Cotinus coggygria Scop. Young fustic; Mordant dye

1167 Rhamnetin Rhamnus infectoria L. Persian berries; Mordant dye

1168 Carminic acid Dactulopius coccus Costa Cochineal; Mordant dye

2029 Lawson Lawsonia inermis L. Henna; Direct dye

2031 Indigo Indigofera tinctoria L. Vat dye

2032 Morin

Chlorophora tinctoria L. / Maclura

tinctoria L. / Morus tinctoria L. Fustic; Mordant dye

2035 Quercetin Quercus velutina Lam. Quercitron; Mordant dye

2036 Rutin Quercus velutina Lam.

2961 Orcein Roccella tinctoria DC Direct dye

2989 Isatin Indigofera tinctoria L.

3177 Haematein Haematoxylum campechianum L. Logwood; Mordant dye

4796 (+) epi catechin Rhus coriaria L. , Rhus cotinus L. Catechu; Mordant dye

4799

(+) catechin

hydrate Rhus coriaria L. , Rhus cotinus L.

(28)

Table 3: Dyes samples structures

Sample number Trivial name Structure

0228 Sandelwood 0670 Purpurin 0894 Kaempferol 0895 Emodin 1153 Brasilin 1154 Maclurin 1156 Turmeric 1164 Berberine

(29)

1166 Fisetin 1167 Rhamnetin 1168 Carminic acid 2029 Lawson 2031 Indigo 2032 Morin 2035 Quercetin 2036 Rutin

(30)

2961 Orcein 2989 Isatin 3177 Haematein 4796 (+) epi catechin 4799 (+) catechin hydrate 7200 Campeche wood

(31)

Object 27

Figure 19: Reversed-phase separation 0228 Sandelwood

Object 30

Figure 20: Overlay UV-spectrum 0228

Object 32

(32)

Object 34

Figure 22: UV-spectrum 0670 t=8,545

Object 36

Figure 23: Reversed-phase separation 0894 kaempferol

Object 38

(33)

Object 40

Figure 25: Reversed-phase separation 0895 emodin

Object 42

Figure 26: UV-spectrum 0895 t=9,525

Object 44

(34)

Object 46

Figure 28: Overlay UV-spectra 1153 brasilin

Object 48

(35)

Object 50

Figure 30: Overlay UV-spectra 1154 maclurin

Object 52

Figure 31: Reversed-phase separation 1156 turmeric

Object 54

(36)

Object 56

Figure 33: Reversed-phase separation 1164 berberin

Object 58

Figure 34: Overlay UV-spectra 1164 berberin

Object 60

(37)

Object 62

Figure 36: Overlay UV-spectra 1166 fisetin

Object 64

Figure 37: Reversed-phase separation 1167 rhamnetin

Object 66

(38)

Object 68

Figure 39: Reversed-phase separation 1168 carminic acid

Object 70

Figure 40: Overlay UC-spectra 1168 carminic acid

Object 73

(39)

Object 76

Figure 42: Overlay UV-spectra 2029 lawson

Object 78

Figure 43: Reversed-phase separation 2031 indigo

Object 80

(40)

Object 82

Figure 45: Reversed-phase separation 2031 indigo in DMSO

Object 84

Figure 46: Overlay UV-spectra 2031 indigo in DMSO

Object 86

(41)

Object 88

Figure 48: Overlay UV-spectra 2032 morin

Object 90

Figure 49: Reversed-phase separation 2035 quercetin

Object 93

(42)

Object 95

Figure 51: Reversed-phase separation 2036 rutin

Object 97

Figure 52: UV-spectrum 2036 rutin t=0,623

Object 99

(43)

Object 102

Figure 54: Overlay UV-spectra 2961 orcein

Object 105

Figure 55: Reversed-phase separation 2989 isatin

Object 107

(44)

Object 109

Figure 57: Reversed-phase separation 3177 haematein

Object 111

Figure 58: Overlay UV-spectra 3177 haematein

Object 113

(45)

Object 115

FIgure 60: Overlay UV-spectra 4796 (+)epi-catechin

Object 118

Figure 61: Reversed-phase separation 4799 (+)catechinhydrate

Object 121

(46)

Object 123

Figure 63: Reversed-phase separation 7200 campeche wood

Object 125

(47)

0 25 50 75 100 125 150 175 200 225 250 275 300 113 100 88 75 63 50 38 25 13 0

Figure 65: Mixture of synthetic dyes 0084 60:40 254 nm

0 25 50 75 100 125 150 175 200 225 250 275 300 113 100 88 75 63 50 38 25 13 0

(48)

0 25 50 75 100 125 150 175 200 225 250 275 300 113 100 88 75 63 50 38 25 13 0

Figure 67: Mixture natural dyes 0086 60:40 254 nm

0 25 50 75 100 125 150 175 200 225 250 275 300 113 100 88 75 63 50 38 25 13 0

(49)

0 25 50 75 100 125 150 175 200 225 250 275 300 113 100 88 75 63 50 38 25 13 0

Figure 69: Mixture of synthetic and natural dyes 60:40 0085 254 nm

0 25 50 75 100 125 150 175 200 225 250 275 300 113 100 88 75 63 50 38 25 13 0

Characterization of natural dyes by comprehensive 2D liquid chromatography

Bachelor Thesis Scheikunde

door

12 juli 2016

Abstract

Samenvatting

Table of contents

1. Introduction

1.1 Dyes

1.2 Liquid Chromatography

k =

t

−

t

t

H ≈ A+

B

u

+

C u

1.3 LCLC

1.4 Separation of dyes

2. Experimental

2.1 Instrumental

2.2 Chemicals

2.3 Analytical conditions

3.1 IP-RPLC

3.2 IEX

3.3 Comprehensive 2D-LC

4. Conclusion

References

Appendix