Mapping the intrinsic viscosity of hyaluronic acid at high conce- ntrations of OH-

(1)

17010

Examensarbete 15 hp Maj 2017

Mapping the intrinsic viscosity of hyaluronic acid at high conce- ntrations of OH-

Mathias Axelsson Anton Lindblad Lovisa Ringström

Johanna Munck af Rosenschöld

Victor Spelling

(2)

Teknisk- naturvetenskaplig fakultet UTH-enheten

Besöksadress:

Ångströmlaboratoriet Lägerhyddsvägen 1 Hus 4, Plan 0 Postadress:

Box 536 751 21 Uppsala Telefon:

018 – 471 30 03 Telefax:

018 – 471 30 00 Hemsida:

http://www.teknat.uu.se/student

Abstract

Mapping the intrinsic viscosity of hyaluronic acid at high concentrations of OH-

Mathias Axelsson, Anton Lindblad, Lovisa Ringström, Johanna Munck af Rosenschöld, Victor Spelling

Hyaluronic acid is commonly used in dermatological fillers in the form of gels. It is established how these gels' firmness is affected by the amount of cross linker and hyaluronic acid respectively. However, the effect of hydroxide ions in solution is rather unknown. This thesis examines how the alkalinity of the solvent affects the intrinsic viscosity of 3 MDa hyaluronic acid by using the method of Ubbelohde capillary viscometry. Sodium hydroxide solutions between 2 and 10 wt% were prepared to study the variation in intrinsic viscosity at concentrations relevant for cross linking (1<wt%). From these respective solutions, four solutions of different mass concentrations of hyaluronic acid were made. The flow time of respective samples were measured between two points in the capillary viscometer in a controlled temperature of 25 °C with an SI Viscoclock to ensure a high accuracy.

From the resulting flow times, the intrinsic viscosity was calculated. The intrinsic viscosity varied between 0,55 and 0,70. The relation between intrinsic viscosity and hydroxide ion concentration had a correlation coefficient r < 0,001. No trend could be ensured as the confidence interval for the intrinsic viscosity at the different concentrations was too large.

Keywords: Alkaline solution, Degradation, Extrapolation, Hyaluronic acid, Intrinsic viscocity, Macromolecules, Polymer, Regression, Sodium hydroxide, Ubbelohde capillary viscometry.

ISSN: 1650-8297, TVE-K17 010 Examinator: Enrico Baraldi Ämnesgranskare: Isak Öhrlund Handledare: Åke Öhrlund

(3)

Acknowledgements

We would like to thank Uppsala University:

Thank you, Jan Bohlin, for all the time, energy and commitment you have invested in us and this project. Thank you for making sure we had the best equipment and information available. And thank you for always finding kind words of encouragement and motivating us from start to finish.

Thank you, Isak Öhrlund, for giving us valuable second opinions, and for guiding us throughout the project.

Thank you, Marit Andersson, for taking the time to help us process our readings and making sure it was performed statistically correct.

Thank you, Christer Elvingsson, for taking the time to discuss the theory be- hind our readings with us and helping us strengthen our arguments.

Thank you, Tim Bowden, for also discussing our results with us and helping to point us in the right direction.

Thank you, Jöns Hilborn, for taking the time to listen to our presentation and discussing our readings with Jan and us.

We would also like to thank Galderma: Thank you, Åke Öhrlund and Mor- gan Karlsson, for always being available to meet for a discussion, providing material and for always showing interest in our work during the project.

(4)

(5)

1. Introduction

1.1 Hyaluronic acid

1.1.1 History of- and applications for Hyaluronic acid

Hyaluronic acid (HA) was discovered and named by Karl Meyer and John Palmer in 1934 when they extracted the molecule from bovine eyes. By that time, they could not determine the exact structure of the molecule. The hy- pothesis of the structure was that it was a polymer contained an amino sugar, a uronic acid and a pentose. Therefore, they suggested to name the molecule hyaluronic acid (hyaloid = vitreous). Even though the structure was not deter- mined the newly found molecule was described as a "polysaccharide of high molecular weight". Determining the structure took almost 20 years [1].

During the 1930’s and the 1940’s other extraction sources were found. These sources include everything from roosterâs comb, human tumour tissue, skin and umbilical cord, and even strains of bacteria. Today, HA is industrially produced from either gene modified bacteria or by extraction from rooster’s comb [1].

The many great properties of hyaluronic acid entail a lot of different areas of application. The first one being eye surgery where it was used as a replacement for vitreous loss in the eye. Other medical applications for HA is to prevent adhesion during surgery and for healing wounds. Furthermore, the elastic and viscous properties of the hyaluronic acid makes it a great polymer for treating arthritis and other joint disorders in humans and even horses [1,2].

Many of its properties makes it a great molecule for application in beauty and cosmetic market. As the polymer binds a lot of water it is very often used as a hydrator in cosmetics and skin care products. The viscous and elastic properties are instead utilized when making cosmetic fillers. By crosslinking HA, gels with desired firmness can be created which can be used for filling out wrinkles and lines. Fillers can also be used to enhance one’s appearance by giving more volume to desired areas, such as lips or cheekbones. As the polymer is a bio polymer, i.e. bio compatible, it is a great polymer to applicate in both medical areas and cosmetics [2].

1.1.2 Chemical properties

Hyaluronic acid (HA) is a bio polysaccharide found i.a. in vertebrate tissues in the extracellular matrix. The name hyaluronic acid can be regarded as slightly

(8)

misleading, since the hyaluronic acid rather is a glycosaminoglycan consist- ing two monosaccharide units; D-glucuronic acid and N-actyl-D-glucosamine, connected byb(1 ! 4) and b(1 ! 3) glucosidic bonds [3], see figure 1.1.

Figure 1.1. Hyaluronic acid.

At physiological conditions HA acts like a salt, i.e. sodium hyaluronan. The strongest acidic group, the carboxylic group, in HA has a pKa value of 2.87 [4]. This means that the carboxylic group will be deprotonated under physiological conditions. The negative charge of the molecule will interact with positively charged ions and molecules, principally Na⁺, within the body [5].

Due to this polyanionic charge and because the HA molecule is not branched, the molecule has a rigid and extended conformation. However, in solutions the HA molecule tend to be in a random coil conformation and this is the conformation that has been found when extracting HA from tissues [1].

HA has an exceptional ability to bind water molecules, where the random coils contain almost 99% bound water. Even at highly diluted solutions the extended molecules get tangled together giving the solution great elasticity and viscosity [5]. There will naturally be a variation of elastic and viscous properties of a HA solutions dependent on the average molecular weight, concentration and the conformation of the polymer. The conformation can differ at different conditions such as pH and temperature. This means that the properties of the polymer are dependent on the conditions in solution [1].

1.1.3 Degradation of Hyaluronic acid

The hyaluronic acid degrades naturally in vivo and, in clinical usage as fillers, last for approximately six to twelve months depending on various circum- stances such as age, lifestyle, individual skin type etc. [6].

To be able to create products of cross linked hyaluronic acid for clinical usage, the knowledge of its physical properties is of interest. During the process of cross linking hyaluronic acid elevated pH is a necessity, though, higher pH cause a degradation of the HA. This results in a decrease in molecular weight of the product in comparison to the raw starting material, which give rise to different physical properties. Prior studies of the degradation of hyaluronic

(9)

acid present a linear behaviour of the degradation, depending on pH respectively of the temperature. In these prior studies the temperature dependent degradation showed greater significance than that of high pH [7].

1.1.4 Hyaluronic acid in aqueous solutions

Studies done on the rheological behaviour of HA in aqueous solutions has proven that there is a pH dependence of the viscosity. At very low and very high pH values the formation of the HA molecule differ from the neutral conformation of HA. At very low pH the molecule stiffens, making the polymer more voluminous. High pH interferes with hydrogen bonding, leading to a smaller volume of the polymer. As the conformation of the molecule changes, the viscosity will most likely change. This means that the viscosity of a solution at a given concentration will vary with pH alterations.

Mutually for these studies there is a distinct increase in the viscosity of solutions at pH < 2.5 and a clear decrease as the pH exceed 11. At physiological pH, the polymer is less sensitive to pH change. At 6.5 < pH < 8.0 there is almost a constant expansion of the HA chains which means the intrinsic viscosity is unchanged [3].

As mentioned earlier the carboxylic group will be deprotonated in neutral solution. Due to intramolecular hydrogen bonds the conformation of the polymer will be rigid. The rigidness increases as the pH reaches pH 2.5. This is due to a protonation of the carboxylic group leading to a formation of an intermolecular H-bond network. The result of this is a viscoelastic, gel-like mass is formed. The viscosity at this pH is therefore higher than at neutral pH.

When decreasing the pH to about 1.6 the polymer is yet again soluble. The explanation for this is that the amido group in addition to the carboxylic group protonates causing an increase in the net charge of the polymer.

In alkali conditions (11 < pH) the viscosity of HA solutions seems to decrease. This is a result of deprotonation of not only the carboxylic group but also hydroxyl groups, which causes the hydrogen bonds to break. Conse- quently, the rotational freedom of the glucosidic bonds increase. This causes a reduction in polymer volume and thereby the solution viscosity will decrease [3,8].

1.2 Galderma AB

As mentioned earlier hyaluronic acid is an eminent bio polymer to use for medical solutions and cosmetics. Because of these great dermatological qualities, Galderma use HA in many of their products. Galderma AB is a global dermatology company that provide medical solutions and treatment for patients

(10)

with different dermatologic issues i.e. acne, rosacea, psoriasis, skin cancer and pigmentary disorders. Another area of expertise is beauty and solutions for aging skin by producing cosmetic fillers. The company was founded by L’Oréal and Nestlé in the year of 1981 and is today a worldwide company with over 6000 employees, a wide product portfolio available in more than 100 countries and 35 wholly-owned affiliates [9].

Galderma Nordic AB, with headquarters in the city of Uppsala, is a result of the take-over of the company Q-Med AB the year of 2011. Q-Med was founded by the entrepreneur Bengt ågerup in 1987 [10]. With a PhD in phys- iology and valuable knowledge and experience of hyaluronic acid from his earlier employers, Pharmacia and Biomatrix, ågerup founded Q-Med to de- velop and commercialise his research of hyaluronic acid [11]. With some small molecular modifications, cross linking of the molecule was made possible and a firm gel was developed- a quality that came to be very useful. The technology is called NASHA, Non-Animal Stabilized hyaluronic Acid, and the patent was granted the year of 1995 [12].

1.3 Aim

The gel in Galdermaâs fillers is constructed by cross linking HA with a cross linker at high concentrations of OH . To be able to produce gels of desired qualities it is of interest to find out how the gel behaves under different conditions. One quality that is usefully adjustable is the firmness which varies according to the conditions during cross-linking.

It is known by the development department of Galderma Nordic, Uppsala, that the resulting firmness of the gel from cross linking will increase, as the concentration of HA and the concentration of cross linker increases. However, it is less known how the concentration of OH affects the molecular volume.

It has been suggested by Galderma that a greater amount of overlap between the molecules during crosslinking may result in a firmer gel. This overlap is affected by the concentration of hyaluronic acid as well as the volume of the polymer. Furthermore, the volume of each polymer is dependent on the surrounding environment, such as the pH. Therefore, it is of high interest to find out how the volume of HA varies at different concentration of OH . One way of measuring the volume of HA is to study the intrinsic viscosity, where an increased intrinsic viscosity corresponds to an increased polymer volume, and vice versa. Thus, the aim of this project is to map how the intrinsic viscosity of 3 MDa hyaluronic acid varies in aqueous solutions with high concentrations of OH .

(11)

2. Theory

2.1 Viscosity

This section is concerned with the general background to viscosity. A more detailed presentation of its application in macromolecules, is given in 2.2 Vis- cosity in solutions of macromolecules.

The viscosity of a fluid is a measurement of the internal resistance that arise in a flux, which is defined as shear stress (t) over rate of shear (g). The relationship is here given by Newton’s equation, eq. 2.1.

h = t

g (2.1)

Viscosity is a state dependent phenomenon that varies with temperature and pressure. In general viscosity increases with increasing temperature and decreasing pressure, for solutions containing polymers [13,14].

A fluid can respond to shear stress in different ways. Based on the behaviour of a fluid it can be classified as Newtonian or non-Newtonian. For a Newto- nian fluid, there is a direct correlation between shear stress and deformation.

This linear dependence means that viscosity remains constant regardless of the shear stress. The response is the opposite for a non-Newtonian fluid, as it can typically behave shear thinning or thickening. In this context, viscosity can either decrease or increase under stress.

For a solution containing macromolecules it is of relevance to compare the viscosity with that of the pure solvent. Relative viscosityh_rel is defined as the ratio between the solution’s viscosity and the pure solvent’s viscosity, see eq.

2.2.

h_rel = h_x

h₀ (2.2)

Specific viscosity hsp is the relative viscosity subtracted by one, given by eq.

2.3.

hsp = h_x

h0 1 (2.3)

In reduced viscosity,h_red is derived from specific viscosity with the only difference that it takes concentration in to account, where the concentration is given in ml/g. This equation can be found in eq. 2.4.

(12)

h_red = h_sp

c (2.4)

Intrinsic viscosity [h] is a parameter used to describe the hydrodynamic properties of a macromolecule [14]. It is defined as 1/concentration; this can be compared with viscosity that has the unit Pa · s. To determine the [h], one can measure the inherent or the specific viscosity. [h] on the other hand is estimated through extrapolation of these values to infinite dilution, with the definition of [h] being according to eq. 2.5.

[h] = lim

c!0

h_sp

c (2.5)

The relationship betweenh and inherent viscosity is given by Huggin’s equation, eq. 2.6, where the specific viscosity is given by eq. 2.7.

h_sp =k₀[h]c + k₁[h]²c²+k₂[h]²c²+ ... (2.6) h_sp

c = [h] + k_H[h]²c (2.7)

Wherehsp,h, kH and C designate specific viscosity, Intrinsic viscosity, Hug- gin’s constant and concentration respectively. The constant k_H is unique for a specific molecule and solvent, in theta-solutions k_H ranges between 0.5 0.7 [15]. In this context theta-solution has favourable polymer-solvent interaction.

The polymer chain becomes flexible and takes the configuration of a random coil [16].

2.2 Viscosity in macromolecules

Highly viscous liquids flow more slowly due to higher resistance compare to low viscous liquids [13]. The size of a molecule has a large effect on the viscosity. Therefore, polymers highly affect the viscosity, even at very low concentrations. Outstretched macromolecules cause higher viscosity compared to coiled up molecules. This is due to outstretched molecules have a larger contact area, which increase the possibility to interact with other molecules.

The interactions between macromolecules appears as overlaps, entanglements, inter-molecular bonds and electrostatic interaction, with one or many other molecules [13].

Coiled up molecules give rise to lower viscous fluids, due to the possibility of interaction with other molecules decrease as the contact area decreases.

Different functional groups dissociate at a range of pH-values in the surrounding environment. Depending on if respective functional group of the polymer are in a dissociated or protonated state, the molecule adopts different charges,

(13)

which give rise to different intra- and/or inter-molecular bonding and electrostatic interactions. A polymer with charged side groups will repel intra molecularly, and the molecule will therefor become more stretched out with an increased volume consequently [16]. Though, in presence of counter ions, charged groups will be shielded and the repelling will therefore decrease. The viscosity of a charged polymer is therefore dependent on both the pH value and the salinity of the solution. Generally, for many polymers the viscosity of a solution will increase as the charge of the polymer increases. The effect of the repellent obviously also affect the inter molecular interactions. Stretched- out polymers have a larger area of interaction which give rise to an increase in viscosity. Though, at the same time, charged molecules repel each other which decreases the viscosity due to increased, possible inter molecular movement [16].

However, other molecular features show a higher importance in the viscosity behaviour of hyaluronic acid, due to the commonly occurring hydrogen bonds.

When going from neutral solution to high OH concentration, hydroxy groups dissociates. This results in breakage of intra- and intermolecular hydrogen bonds, giving the molecule free rotation at the glucosidic bonding. This causes the molecule to a have less rigidity which increases the ability to coil up to a smaller volume. Even though there will be more charge within the molecule leading to repelling, the effect of breaking the hydrogen bonds is of greater impact [1,3].

2.3 Viscometry

Two of the more general methods used to determine the viscosity of a fluid are capillary viscometry and rotating rheometers. These two are used in different situations as they measure different things, especially rheometry, which has a wider range of applications. In capillary viscometry, time is measured and viscosity can be calculated. In rheometry, one can apply a controlled force or strain to measure the responding force or strain of the liquid; and this force or strain can be varied stepwise and thus give more information from one single run. This makes rheometry suitable to measure non-Newtonian fluids [17]. Rheometry can also be used for measurements on solid and semi-solid materials.

2.3.1 Rheometry

In rheometry the bob measures at a controlled torque or speed, and rotational or oscillating movement. What setup is used depends on the substance and/or what one wants information about, e.g. the viscosity of ketchup is best measured with controlled shear-rate test and the movement should be rotational

(14)

instead of oscillating because ketchup is a non-Newtonian fluid. With this test, the shear-rate could be changed stepwise and the changes in viscosity could be examined [18].

2.3.2 Capillary viscometry

When using capillary viscometry there are different instruments to choose from. In this study the chosen capillary was a Ubbelohde viscometer, normal form (ASTM), Capillary No.I. The principle of Ubbelohde capillary is simple:

Time is measured for an arbitrary volume of given solution to flow through a capillary of a specific length and diameter. The flow time can be registered manually or automatically. Automatic registration with a Schott Instrument (Hoffheim, Germany) ViscoClock system is a clear advantage over manual registration as the manufacturer [19] states the measurement accuracy to be

±0,01 s; with this automatic system, subject measurement error is eliminated.

To conduct measurement in an Ubbelohde capillary, specific requirements need to be met. First, temperature needs to be controlled; by praxis opera- tions are carried out in 25 C for capillary viscometry. Second, concentration is another critical factor. At high concentration, there are multiple molecular interactions between adjacent molecules. This is an associative phenomenon of HA [1], which means solutions in this project will assume non-ideal states.

Both the reduced viscosity and inherent viscosity are concentration dependent [1]. To calculate these values the concentration needs to be defined with a high accuracy. Intrinsic viscosity on the other hand is defined at zero concentration.

This gives a theoretical value for the polymers volume that is independent of the concentration. Therefore, this value is suitable for comparison.

Capillary size is of importance when measuring as it must operate well for given solvent and solute. By changing to a capillary with a different diameter, it is possible to change the flow rate. It is essential that have the appropriate capillary size to get a reliable measurement. In this project, the solvent used was water and therefore the capillary size needs to be large enough to avoid capillary action propelling the sample and thereby affecting the flow time.

The solute properties are also of importance when choosing capillary, as in this project. HA has a non-Newtonian fluid behaviour. Under stress it behaves shear-thinning. Hereby a problem arises since the gravitational force, in the capillary, may affect. In a publication [14] by Steve Harding, a theory is being presented that a non-Newtonian fluid can behave as a Newtonian fluid at low flow rate.

(15)

2.4 Data analysis

To calculate the intrinsic viscosity from the flow times, several calculations must be made. At first one calculates is the relative viscosity, eq. 2.2. In this experiment, the density of the solution and of HA were considered equal, giving the reduced form of eq. 2.8 as eq. 2.9, which is used in the calculations.

h_rel = t_x t₀ ·rx

r₀ (2.8)

h_rel = tx

t₀ (2.9)

The second step is to calculate the specific viscosity from the relative viscosity, using eq 2.10.

h_sp = t_x t₀

t₀ (2.10)

In order to estimate the intrinsic viscosity, [h], reduced viscosity h_red 2.11 is plotted against mass procent HA (see figure 4.1.)

h_red = t_x t₀

ct₀ (2.11)

A regression line is fitted to the scatter plot and from the intercept on the ordinate axis, the intrinsic viscosity is acquired[20]. In order to accurately estimate the true intrinsic viscosity, the uncertainty of the intercept needs to be coincided. Correlation coefficient and regression analysis are two good methods that gives complementing information to the analysis.

Correlation coefficient

All the intrinsic viscosity values were plotted against their respective wt%

NaOH, allowing the graph to display the mapping that was sought. A correlation analysis was made of the graph with eq. 2.17 [21].

Correlation coefficient r describes the linear ’co-relations’ between two quan- titative variables. Here, designated X and Y. It indicates on both direction and strength in the relationship. For a high correlation coefficient, the variability is well explained. This is shown by a r-value close to -1 or 1. The sign simply indicates whether the relationship is positive or negative. If Y increases with X, the relationship is said to be positive. The opposite is true for a negative relationship.

Correlation coefficient can be derived from the covariance Cov(X,Y).

Cov(Y,X) = Â[(x_i x)(y_i y)]

n 1 (2.12)

(16)

Covariance is closely related to correlation coefficient, with one mayor difference. Covariance can not be used to estimate the strength in the measurement.

Instead data needs to be standardised, by dividing them with the standard deviation in x and y. Here follows a example of X.

z_i= (x x_i)

s_x (2.13)

where

s_x=

r 1

n 1

Â

^(xⁱ ^x)² ^(2.14)

is the standard deviation in X. The same is preformed for Y. Finally, correlation coefficient is given by covariance.

r = Cov(X,Y )

s_xs_y (2.15)

Corelation coeffiecient can also be written as r = 1

n 1

Â

^(xⁱ _s_x^x)² ^(yⁱ _s_y^y)² ^(2.16)

r = Â[(xi x)(y_i y)]

p[Â(xi x)²][Â(yi y)²] (2.17) .

Regression analysis

Linear regression analysis can be used in descriptive purpose, as with correlation coefficient. In addition to this, regression analysis can be used to predict.

For a given value of x, the correspondent y-value can be calculated with a associated confidence interval. This makes it possible to extrapolate a trend line to the intercept on the y-axis. From this intercept the intrinsic viscosity is estimated with some uncertainty given by the confidence interval.

Different models are used in linear regression analysis eg. least square method and maximum likelihood estimation etc. Fully developed mathematical de- scription could be found in a statistical textbook.

R-square

To statistically secure that calculated values are close to the regression line, one can calculate an R² value (goodness of fit). If R² equal to 0 (0 %), the model doesn’t explain the variability of the points along the regression. If R² is equal to 1 (100 %), the model does explain all the variability of the points.

In other words, the higher R² value, more points variability will be explained by the regression model. [22]

(17)

3. Method

3.1 Choice of method

To obtain results precise enough to be able to calculate the relative viscosity, and thereafter the intrinsic viscosity, a method with high precision was sought.

Capillary viscometry was chosen as method, seeing as an available SI Visco- clock had an accuracy of 0.01 % for time measurements, which was judged to be the most superior method. The SI Viscoclock was connected to an Ubbe- lohde capillary, making it possible to measure the flow time automatically, which provides more accurate results compare to manual clocking. Beyond being easily accessible it is also an economically beneficial method for this type of study.

A capillary size of 0.58 mm was chosen with the properties of water as a solvent and HA as a solute considered. The running terms for capillary viscometry of measuring at 25 C was facilitated in a temperature-regulated tank Heto-Holten A/S (Allerød, Denmark) type 21DT-2.

3.2 Procedure of method

The procedure of the method is summarised in a flow chart below, presented in figure 3.1.

(18)

Lye

dilution Reference run

Hyaluronic acid dilution

40 minutes of stirring

10+2 minutes of

tempering

Run thrice in capillary

Clean and dry the capillary

Repeat

Figure 3.1. Method illustrated as flowchart

3.2.1 Solutions

To outline the variation in intrinsic viscosity at high [OH ], NaOH solutions of 2, 4, 6, 7, 8, 9 and 10 wt% respectively were prepared. For each concentration of NaOH a water solution of approximately 200 mL was prepared in a round bottom flask, by first weighing milli-Q water followed by the required mass of NaOH pellets to achieve the desired mass concentration. From this solution four Hyaluronic acid solutions, of different mass concentrations, were made.

An example is given in Table 3.1, which presents the compositions of the four different HA-solutions at the NaOH-concentration of 2 wt%.

Table 3.1. Dilution series for the four HA solutions at 2 wt% of NaOH HA (g) NaOH-solution (g) [HA] (mg/g)

C1 1.40 32.60 0.800

C2 1.05 33.95 0.600

C3 0.700 34.30 0.400

C4 0.350 34.65 0.200

Each solution was kept under magnetic stirring for 40 minutes, followed by 10 minutes of tempering in 25 C thermostatic bath. Afterwards 17 mL of tempered solution was measured in a measuring cylinder from which it was poured into a plastic syringe and therefrom dosed into the Ubbelohde viscometer through a PVDF membrane filter. Before starting measurements the

(19)

dosed viscometer, attached to the SI Viscoclock, was intentionally left for 2 minutes of tempering in the 25 C thermostatic bath.

3.2.2 Measurements

For each measurement of the HA-solutions, the flow time in the capillary was clocked with the SI Viscoclock. Each solution was repeatable measured thrice and a mean time was calculated. In addition to this a reference flow time was also clocked thrice, from the NaOH-solution without any HA, and later used in the calculations. The time in between the repeated measurements of each sample was kept as short as possible to minimise the time dependent degradation of the polymer while in the NaOH-solution. In between the different solutions the capillary was washed out with distilled water, followed by KOH soap and acetone so that it could easily be dried with N₂ gas.

All the data processing and analysis processing was done with Microsoft Of- fice Excel 2016.

3.3 Method Development

Before the method was established, pilot-laboratorial work was made to de- terminate and predict difficulties throughout the process. The same laboratory equipment was used as in the latter experiments. The measurement of each solution was also made repeatedly thrice. Though, the concentrations of NaOH varied compare to the solutions of the latter experiments (3, 5, 7, 8.5 and 10 wt% NaOH respectively).

During the pilot project, observations made it clear that the time consistency during the sample preparations were of importance due to the degradation of HA. The importance of tempering each sample for an equal amount of time was also evaluated. The experience from the pilot project resulted in a higher knowledge of how important it was to keep the consistency throughout the laboratorial work. From this knowledge, the decision to always stir and temper each sample for an equal amount of time in latter experiments was made. The results from the pilot project was valued not to be credible enough to present and will therefore be disregarded.

To satisfy Galderma’s acceptance requirements of measurements of the viscosity of HA, the quota between the mean flowtime of the highest concentration (C1) and the reference mean flow time was to fall within the range 1.6 and 1.8. By varying the amount of HA in the solutions, the acquired range could be accomplished. During the pilot-laboratory work it was found that the searched range was implemented in the solution of 3 wt% NaOH at HA- concentrations of 0.8 [mg/g] for C1, respectively HA-concentrations of 1.0 [mg/g] for solutions between 5-8.5 wt% NaOH.

(20)

4. Results

In this chapter our laboratory findings are presented. The results are given in a chronological order, which ultimately leads to the intrinsic viscosity. Since the procedure was the same for 2, 4, 6, 7, 8, 9 and 10 wt%, a detailed explanation is only given for 2wt%. The measured values for the others can be found in appendix 1.

The exact mass concentration of the 2 wt% NaOH solution is shown in table 4.1. The table also shows the amount of water and NaOH(s) that were weighed to create the solution.

Table 4.1. The weighed masses of water and NaOH for the 2 wt% solution and its exact concentration value.

Water (g) 185,16

NaOH (g) 3,7793

Concetnration (wt%) 2,000

Four samples with different masses of Hyaluronic acid (HA) were prepared.

The amount of HA and NaOH(aq) needed, and their resulting concentration are presented in table 4.2.

Table 4.2. The weighed masses of HA and 2 wt% NaOH(aq) that were mixed and the resulting concentrations of the samples used in the experiments.

Cx HA (g) NaOH_(aq) (g) Concentration (mg/g)

C1 1,403 32,6829 0,8232

C2 1,058 33,9505 0,6044

C3 0,7037 34,3026 0,4020

C4 0,352 34,6503 0,2011

The results of the 2 wt% NaOH solution measurements are found in tables 4.3 and 4.4. The first one presenting the flow time of the different concentrations and reference sample, along with the mean values, standard deviations and relative standard deviations. The latter presents the relative, specific and reduced viscosities calculated from the flow times and concentrations (in table 4.2).

(21)

Table 4.3. Observed flow times in the Ubbelohde capillarry of the reference and the samples, with a calculated mean, standard deviation and relative standard deviation.

For each run, the flow time are clocked three times.

Flow time (s)

Ref C1 C2 C3 C4

110.32 189.59 161.44 140.73 124.48 110.16 188.4 160.98 140.54 124.44 110.46 187.47 160.65 140.70 124.39 Mean time (s) 110.31 188.49 161.02 140.55 124.44 Std.dev. (s) 0.1501 1.063 0.3968 0.1801 0.04509 Relativ std.dev. (%) 0.1361 0.5638 0.2464 0.1281 0.03624

Table 4.4. The relative, specific and reduced viscosities of samples of 2 wt% NaOH solution. These values are calculated, based on the rawdata presented in table 4.3

h_rel hsp h_red C1 1,709 0,7086 0,8608 C2 1,460 0,4597 0,7605 C3 1,274 0,2741 0,6817 C4 1,128 0,1280 0,6366

The calculated reduced viscosities were plotted against the different HA concentrations, with the reduced viscosity on the y-axis and HA concentration on the x-axis. This graph can be seen in figure 4.1.

(22)

Figure 4.1. Graph showing the resulting reduced viscosities [cm3/g] on the y-axis and the HA concentrations [mg/g] on the x-axis for measurements at 2 wt% NaOH solution.

All the intrinsic viscosities were obtained at the intercept with the y-axis in the graphs shown in figure 4.2.

(23)

Figure 4.2. All the graphs where the reduced viscosity is plotted against the NaOH solutions (a-2, b,c-4, d-6, e-7, f-8, g-9, h-10 wt%) and where the intrinsic viscosities are obtained at the intercept with the y-axis.

The intrinsic viscosities of HA at different NaOH concentrations are presented in table 4.5, along with the corresponding NaOH concentrations and a 95,0%

confidence interval’s upper and lower limits for the intrinsic viscosity values.

The table’s values are calculated from the graphs in figure 4.2.

A graph showing the NaOH-concentration dependence of the intrinsic viscosity is found in figure 4.2. The error bars in the graph are the confidence interval for each point.

(24)

Table 4.5. All the estimated intrinsic viscosities, their upper and lower confidence interval limits, their corresponding NaOH solution concentrations, p-values and R- square values.

NaOHaq(wt%) IV (cm³/g) Upper 95.0% CI Lower 95.0% CI p-value t-value R-square

2.000 0.5499 0.6403 0.4595 0.01051* 26.17 0.9791

4.006 0.6522 0.8105 0.4938 0.1583 17.72 0.7084

4.159 0.6537 0.7066 0.6009 0.02436* 53.19 0.9519

6.008 0.6091 0.6787 0.5396 0.02734* 37.68 0.9461

6.951 0.6136 0.7191 0.508 0.08171 25.01 0.8433

8.005 0.689 0.7897 0.5884 0.8588 29.45 0.01995

8.998 0.5826 0.6461 0.5191 0.01917* 39.48 0.9620

10.00 0.5743 0.6485 0.5191 0.04589* 33.29 0.9103

*Significant on a 5% level.

Figure 4.3. The intrinsic viscosity’s NaOH dependence are illustrated in the scatter plot with the confidence interval as error bars. Intrinsic viscosity [cm3/g] on the y-axis and NaOH solution [wt%] on the x-axis.

The relationship between IV and massprocent NaOH was evauated. Corre- lation coefficient was calculated of the estimated mean intrinsic viscosity, to r = 0.0006265.

(25)

5. Discussion

5.1 Data Analysis

As can be seen in figure 4.3 the error bars have a wide range, meaning a wide confidence interval. With such high uncertainty a final conclusion of a trend can not be made. As of now a trend line could be drawn horizontally within the confidence interval of all the points, which would indicate that the intrinsic viscosity does not change at all with the NaOH concentration. To draw any conclusions, this uncertainty needs to decrease. Even though two separate measurements were made around 4 wt% it might be a random coincident that they were roughly the same, seeing as one of them had a significantly larger confidence interval.

If the degradation of the polymer could be avoided, the standard deviation (std.dev.) of the flow time would improve as the systematic loss in seconds between the measurements would not occur. This in turn should have resulted in a narrower spread around the true mean. If the std.dev. could be decreased the confidence interval would improve, e.g. at C2 for 7 wt% the mean difference between t₀ and t_x is 77 [s] and t_x has a std.dev. of 1,2 [s]. If then the std.dev. was added to the mean and then divided with the mean, the result would be 1.5 % increase. This 1.5 % difference in the mean of measured time will follow through the rest of the calculations being made and giving them at least the same uncertainty.

For the graphs a-h, in figure 4.2, 95.0 % confidence intervals have been calculated and tabulated in table 4.5. These confidence intervals refers only to the intercept with the y-axis. No confidence lines around the regression have been visualised.

Overall, the calculated p-values are < 0.05, but the first 4 wt%, 7 wt% and the 8 wt% shows a higher value and they all have a lower R-square value than the ones with p<0.05. For these three measurement, Y does not depend on X in a linear way. The intrinsic viscosity calculated from these three measurement, are weak estimations.

Why the first regression for 4 wt% NaOH show poor values and confidence interval might be a result of problems with the thermostatic bath during runs.

C2 and C3 ran at 26 C, and not 25 C where C1 and C4 were run at. This could have resulted in a viscosity difference for those samples. Why 7 wt% values p > 0.05 is unknown. There might have been something wrong with C3 if one study the 7 wt% graph in figure 4.2. C3 is not as close to the regression line

(26)

as the other and causes the R-square to decrease. Why the 8 wt% values are poor we do not know. The observed flow times are without any noteworthy oddities but the reduced viscosities do not follow a linear trend like the others regressions do. This regression has an R-square value of 0.01995, meaning that the model is bad at explaining the variability of the points. Said results makes these graphs very questionable and thus also the extrapolation resulting in the intrinsic viscosity used in the final results. With all having a p>0.05 we can not say that Y depends on X.

5.2 HA in solutions

Data in documents provided by Galderma presents that the intrinsic viscosity for this batch of HA in neutral solutions is 2.9 cm³/g. As mentioned earlier the intrinsic viscosity should decrease once solution reaches pH < 11. In this project, the pH ranges from 13.7 < pH < 14.4 and the estimated intrinsic viscosity ranges between 0.55 and 0.69 cm³/g. Thus, there is a great decrease of the intrinsic viscosity relative that at pH < 11, as predicted. This indicates that there has been a dissociation of the hydrogen within the hydroxy groups, leading to a disruption of the hydrogen bond network. This results in shrinkage of the polymers volume since there is free rotation of glucosidic linkages.

Smaller volume of the tangled polymer leads to a decreased chance of overlap with other HA molecules within the solution.

Even though the interval of the wt% is rather large, the corresponding interval of pH is very small. This is because the scale of pH is logarithmic. It is therefore reasonable that the intrinsic viscosity would be approximately the same for all samples due to lack of variation in the alkalinity. The results found show that there are two peaks at 4 wt% and 8 wt% respectively, albeit these results are not reliable since the confidence interval is very wide, as mentioned earlier. It is not likely that the true value of intrinsic viscosity would vary the way it seems with the diverging values of 4 and 8 wt%, since the interval of pH is very narrow. These two diverging values are probably a result of random experimental errors.

Furthermore, if the resulting values of 4 and 8 wt% were to be ignored, a slightly decrease in the intrinsic viscosity could potentially be observed. This could possibly be explained by the shielding effect of [Na⁺]. By increasing the concentration of NaOH more cations will be available in the solution, in- tensifying this salt effect. The net charge of the polymer will thereby decrease without it regaining the ability to hydrogen bond. This will in turn lead to less intra molecular repelling, resulting in a smaller volume. This theory is not confirmed but could be tested by adding NaCl to the solution to see whether the viscosity would decrease even more. A decrease would strongly indicate

(27)

that the polymer is indeed affected by this salt effect. This is therefore something that could be investigated in the future.

5.3 Degradation

Previous research show a degradation of HA when it is exposed to high tem- peratures or strongly alkaline solution. The degradation shows a linear behaviour, where the temperature is contributing to a higher extent compared to the alkalinity. Degradation of the hyaluronic acid leads to shorter length of the molecules, which in turn lowers the viscosity. The degradation is most likely the reason why the three flow times for each solution decreases in descending order. This behaviour is observed in all samples, except from one (sample C2, 4 wt%) which most likely is an error due to contamination of the equipment.

It is difficult to say what exact effect the degradation has on the resulting intrinsic viscosity. It is probably the case that the entire graph is displaced vertically since all the observed reduced viscosity values are lower than the real ones; this would result in a lower calculated [h] compared to a scenario where degradation does not occur. This displacement would not be propor- tional throughout the four points of the regression. Since C1 spends more time in the capillary while being measured, the significance of the degradation is greater than it is for C4.

Since the degradation is unavoidable, method developments were made to ensure a small time variation, and thereby variation in degradation, as possible between samples. By working under a consistent time frame during the prepa- ration of each sample, the time between the mixing of HA with the NaOH solution, and the first run in the capillary is kept the same. Since every sample is measured thrice, the procedure between measurements has been made as efficient as possible to avoid further time differences.

However, since the flow time differs in between samples, as a result of change in concentration and thereby also in viscosity, the starting time of runs two and three varied. Higher concentration of HA and NaOH in the samples increased the flow time. Even though the second and third run was started as soon as the previous run was finished different extents of delays occurred, giving rise to a slight uncertainty throughout the method. One other possible method that would minimise the effect of the degradation is the idea of doing triplets of each sample that is being measured only once. However, this procedure would be much more time consuming and acquire three times more work material.

5.4 Shear-thinning

The capillary viscometer has a time specific operating range that is applied to Newtonian fluids. Though, uncertainties can occur when measuring non-

(28)

Newtonian fluids. Fluids that are shear-thinning such as the hyaluronic acid can, if the capillary viscometer radius is too large, flow disproportionally fast which will entail to incorrect results of the intrinsic viscosity. Non-Newtonian fluids can however be approximated as Newtonian ones if the flow time is suf- ficiently low. It is therefore important to use a capillary with a suitable radius for the solute to avoid deceiving results. However, it has not been possible to investigate if shear-thinning of the hyaluronic acid take place in the capillary used, which is presenting an experimental error to consider [14].

5.5 Future research

Our results found clearly indicate that a degradation of the polymer occurs during measurements as there is a consistent reduction in flow time. According to a publication by Steve Harding [14], capillary viscometry should not be used for a polymer that degrades during measurements. However, capillary viscometry is an established and well-tried method used to measure viscosity [13], and combined with a SI Viscoclock high accuracy is to be achieved.

It is also an easily accessible and economically beneficial method. Though, a greater amount of measurements on each wt% must be made to ensure a more reliable result and to obtain a satisfactory confidential interval. Anyhow, if further investigations of an alternative method would be of interest, one suggestion would be to use light scatter spectrophotometry, since it improves the time aspect and therefore the significance of degradation.

One very important thing to keep in mind is that the temperature has a big impact on the degradation [7]. It is therefore highly important to keep the temperature stable. Furthermore, as the degradation of the polymer is time dependent, it is of interest to continue making sure that each solution spends the same amount of time in the process of stirring respectively tempering, as well as in the capillary.

In further investigations it should, as a suggestion, be of interest to do the study at a larger interval of pH, to be able to see at which exact pH the dissociation of hydroxyl groups occurs. Lastly it would be interesting to add NaCl to the solution to establish whether there is a common ion effect in the solution or not.

(29)

6. Conclusions

Given the results, no trend can be observed in the intrinsic viscosity between 2 and 10 wt% NaOH. The degradation is great enough to be noticeable during the tests, which causes an uncertainty in the results. Using an Ubbelohde capillary viscometer for the experiment has been difficult to use, and a better suited method may be found. If there is a variation in molecular volume of HA depending on different concentration of OH , it could not be detected in our study due to the uncertainty of the results. However, one can with certainty say that the intrinsic viscosity, and thereby also the molecular volume, is lower at these higher concentrations of NaOH than in neutral solution.

(30)

References

[1] Garg H G, Hales C A, Hales C A. Chemistry and Biology of Hyaluronan [Internet]. Amsterdam: Elsevier Science; 2004. [cited 2017 May 15]. Avail- able from: ProQuest Ebook Central

[2] NE.se [Internet], Nationalencyklopedin. Hyaluronsyra. [cited 2017 May 18]. Available from:

http://www.ne.se/uppslagsverk/encyklopedi/lång/hyaluronsyra

[3] Cowman M K, Schmidt T A. , Raghavan P, Stecco A. Viscoelastic Prop- erties of Hyaluronan in Physiological Conditions [version 1; referees: 2 ap- proved]. F1000Res. 2015 Aug 25;(4)622:5-6

[4] Drugbank. Hyaluronic acid [Internet]. Place unknown. [cited 2017 May 19] Available from: https://www.drugbank.ca/drugs/DB08818

[5] NE.se [Internet]. Nationalencyklopedin. Hyaluronsyra. [cited 2017 May 17]. Available from:

http://www.ne.se/uppslagsverk/encyklopedi/lå/hyaluronsyra

[6] Galderma. Frågor och svar Restylane [Internet]. Galderma; [cited 2017 May 16]. Available from: http://www.restylane.se/behandlingar/Vanliga-fragor/

[7] Fagerström Troncoso J, Idjbara A, Kalrsson I, Lekander M, Lindgren T, Ström S. A kinetic study of the degradation of hyaluronic acid at high concentrations of sodium hydroxide [Bachelor’s thesis on the Internet]. Uppsala: Up- psala Universitet; 2016 [cited 16 May 2017]. Available from: http://uu.diva- portal.org/smash/get/diva2:954372/FULLTEXT01.pdf

[8] Gatej I, Popa M, Rinauo M. Role of the pH on Hyaluronan Behavior in Aqueous Solution. BIOMACROMOLECULES:2015(6);61-67.

[9] Galderma. About Galderma, a Dermatology Company [Internet]. Lau- sanne: Galderma S.A; [cited 2017 May 15]. Available from:

http://www.galderma.com/About-Galderma

[10] Galderma. Om Galderma [Internet]. Uppsala: Galderma Nordic AB;

[cited 2017 May 15]. Available from: http://www.galderma.se/Om-Galderma [11] Entreprenörskapsforum. Entreprenör, forskare, uppfinnare, filantrop...

[Internet]. Stockholm: Entreprenörskapsforum; [cited 2017 May 15]. Avail- able from:http://entreprenorskapsforum.se/natverk/globaliseringsforum/bengt- agerup/entreprenor-forskare-uppfinnare-filantrop%E2%80%A6/

(31)

[12] Mellgren, E. Han tämjde molekylen [Internet]. Stockholm: Ny Teknik;

2010 [Updated 2010 October 12; cited 2017 May 16] Available from:

http://www.nyteknik.se/innovation/han-tamjde-molekylen-6424344

[13] Atkins P, de Paula J. Atkins’ Physical chemistry. 10 ed. Oxford: Oxford University Press, 2014.

[14] Harding S.E. The intrinsic viscosity of biological macromolecules. Progress in meassurment, interpitation and application to structure in dilute solutions.

Prog. Biophys. Molec. Biol. 1997; 68: pp. 207-262.

[15] Sakai T. Huggins constant k’ for flexible chain polymers. J. Polym. Sci., Part B: Polym. Phys. 1968;8(6) pp. 1535-1549.

[16] Piculell L. "Polymerlösningar och geler". Handout. Lund university.

Uppsala. 1997. Print.

[17] ATA Scientific Instrument [Internet]. [Place unknown]: ATA Scientific Instrument; [Publication unknown]; An introduction to viscosity and rheology; 2010 Feb 10 [cited 2017 May 15];[about 2 screens]. Available from:

https://www.atascientific.com.au/an-introduction-to-viscosity-and-rheology/

[18] Anton Paar blog [Internet]. [Place unknown]: Jürgen Utz; 2016 June 9.

Rheometer âWhat do They Measure and How are They Set Up?; 2016 June 9 [cited 2017 May 17]; [about 3 screens]. Available from: http://blog.anton- paar.com/rheometers-what-do-they-measure-and-how-are-they-set-up/

[19] Schott-Geräte [Internet]. Allemagne: Schott-GerÃte GmbH; [Publica- tion unknown]; Operating Instructions Viscosity Measuring Unit ViscoClock;

1997 May 20 [cited 2017 May 17].

[20] Harding S.E (2013) Intrinsic viscosity, in Roberts G.C.K. (ed) Encyclo- pedia of Biophysics, pp. 1123-1129, Berlin: Springer Verlag

[21] Miller N J, Miller C J. Statistic and Chemometrics for Analytical Chem- istry. 6 ed. Harlow: Prentice Hall; 2010. 278p.

[22] The Minitab Blog [Internet]. [Place unknown]: Jim Frost; [2013 May 30];

Regression Analysis: How do I Interpret R-square and Assess the Goodness- of-Fit?; 2013 May 30 [cited 2017 May 31]; [about 4 screens]. Available from:

http://blog.minitab.com/blog/adventures-in-statistics-2/regression-analysis-how- do-i-interpret-r-squared-and-assess-the-goodness-of-fit

(32)

Appendix 1

Table 6.1. 4 wt% (1st): the weighed masses of water, NaOH(g), HA and NaOH_(aq), the concentration of NaOH_(aq) (wt%) and HA (mg/g) and the values of hrel,hsp and hred (cm³/g).

4 wt% NaOHaq(1st)

H20(l) (g) NaOH(g)(g) Conc. (wt%)

182.94 7.939 4.1595

HA (g) NaOH(aq)(g) Conc. HA (mg/g) hrel hsp hred(cm³/g)

C1 1.748 33.34 0.9965 1.751 0.7515 0.7542

C2 1.431 31.39 0.6237 1.665 0.6649 0.7626

C3 1.093 33.96 0.6237 1.434 0.4336 0.6952

C4 0.6977 34.36 0.3981 1.281 0.281 0.706

Table 6.2. 4 wt% (2nd): the weighed masses of water, NaOH(g), HA and NaOH_(aq), the concentration of NaOH_(aq) (wt%) and HA (mg/g) and the values of hrel,hsp and hred (cm³/g).

4 wt% NaOHaq(2nd)

192.06 8.016 4.007

C1 1.757 33.3 1.002 1.755 0.7551 0.7533

C2 1.405 33.62 0.8021 1.599 0.5987 0.7464

C3 1.051 33.94 0.6008 1.431 0.4306 0.7167

C4 0.7033 34.33 0.4016 1.278 0.2783 0.6931

Table 6.3. 6 wt%: the weighed masses of water, NaOH(g), HA and NaOH_(aq), the concentration of NaOH_(aq) (wt%) and HA (mg/g) and the values ofhrel,hspandhred

(cm³/g).

6 wt% NaOHaq

196.83 12.58 6.008

HA (g) NaOH_(aq)(g) Conc. HA (mg/g) hrel hsp hred(cm³/g)

C1 1.75 33.25 0.9998 1.746 0.7458 0.7459

C2 1.397 33.41 0.8028 1.566 0.5661 0.7051

C3 1.053 33.94 0.6017 1.410 0.4104 0.682

C4 0.7063 33.40 0.4141 1.277 0.2771 0.6691

(33)

(cm³/g).

7 wt% NaOHaq

H20_(l) (g) NaOH_(g)(g) Conc. (wt%)

185.87 13.89 6.951

HA (g) NaOH_(aq)(g) Conc. HA (mg/g) h_rel h_sp h_red(cm³/g)

C1 1.746 33.59 0.988 1.717 0.7176 0.7263

C2 1.402 33.94 0.7935 1.559 0.5591 0.7047

C3 1.049 33.99 0.5988 1.397 0.3966 0.6622

C4 0.6957 34.31 0.3975 1.265 0.2655 0.6679

(cm³/g).

8 wt% NaOHaq

199.38 17.3508 8.006

C1 1.751 33.26 1.001 1.706 0.7058 0.7054

C2 1.398 33.61 0.7935 1.547 0.5467 0.6845

C3 1.06 33.95 0.6054 1.413 0.4132 0.6826

C4 0.6867 34.3 0.3925 1.275 0.2754 0.7016

(cm³/g).

9 wt% NaOHaq

182.05 18.00 8.998

HA (g) NaOH_(aq)(g) Conc. HA (mg/g) hrel hsp hred(cm³/g)

C1 1.752 33.25 1.001 1.733 0.7325 0.7318

C2 1.412 33.59 0.8069 1.556 0.5563 0.6894

C3 1.048 33.97 0.5987 1.398 0.3976 0.6641

C4 0.7092 34.31 0.4051 1.262 0.2615 0.6456

(34)

(cm³/g).

10 wt% NaOHaq

H20_(l) (g) NaOH_(g)(g) Conc. (wt%)

200.74 22.33 10.01

HA (g) NaOH_(aq)(g) Conc. HA (mg/g) h_rel h_sp h_red(cm³/g)

C1 1.746 33.25 0.9978 1.678 0.6784 0.6799

C2 1.407 33.6 0.8038 1.525 0.5248 0.6529

C3 1.054 33.95 0.6022 1.391 0.3913 0.6498

C4 0.7047 34.31 0.4025 1.246 0.2459 0.6109

Table 6.8. The measured times for all the experiments.

2 wt% 4 wt%

(1st) 4 wt%

(2nd) 6 wt% 7 wt% 8 wt % 9 wt% 10 wt%

t (s) t (s) t (s) t (s) t (s) t (s) t (s) t (s)

110.32 118.7 116.56 129.78 136.38 144.51 150.13 163.86

Ref. 110.16 118.53 116.47 129.55 136.35 144.47 149.97 163.86

110.46 118.6 116.36 129.5 136.28 144.48 149.92 163.88

189.59 209.63 205.64 228.07 236.13 248.58 262.08 277.61

C1 188.4 206.8 204.37 226.16 234.1 246.48 259.7 275.06

187.47 205.42 203.21 224.59 232.3 244.32 257.88 272.43

161.44 198.45 186.55 204.21 213.82 224.76 235.04 251.6

C2 160.98 197.54 187.5 202.85 212.47 223.48 233.5 249.87

160.65 196.44 184.51 201.87 211.43 222.18 231.84 248.13

140.73 170.6 167.16 183.79 191.24 205.19 211.01 229.41

C3 140.54 170.12 166.6 182.68 190.35 204.19 209.51 227.92

140.37 169.41 166.07 181.93 189.62 203.18 208.42 226.64

124.48 153.13 149.22 166.05 173.08 185.13 190.08 204.71

C4 124.44 151.51 148.84 165.51 172.51 184.32 189.02 204.23

124.39 151.15 148.57 165.02 172.00 183.37 188.61 203.55

Mapping the intrinsic viscosity of hyaluronic acid at high conce- ntrations of OH-