Protocells: Promising nanocarriers for on-demand drug ...uchi.vscht.cz/uploads/svk/E1.pdf ·...

SVK 2018, VŠCHT Praha, Ústav chemického inženýrství

Protocells: Promising nanocarriers for on-demand drug delivery

Martin Balouch

školitel: prof. Ing. František Štěpánek, Ph.D.

školitel – specialista: Ing. Marek Šoltys

Keywords: Protocells, liposomes, silica

Introduction

Liposomes, lipid bilayer vesicles, are the subject of intensive research for many possible

applications in drug delivery. Active pharmaceutical ingredients (APIs) are encapsulated directly in

the lipid bilayer or in the aqueous solution inside of the liposome capsule. The liposome prevents a

premature release of the API and also protects it from the surrounding environment. To date there are

more than 10 approved formulations on the market and more than 20 other formulations are in the

stage of clinical trials [1] that utilize liposomes. However, not all APIs are suitable to be formulated

using liposomes due to the water solubility of the APIs, leakage through the bilayer or disruptive

interaction with the lipidic membranes of the liposomes. One of the possible solutions for these

problems is to first adsorb the API into a porous particle and then encapsulate this particle itself into

the liposome. Such formed nanocomposites are called protocells. Protocells were firstly prepared by

Ashley and Cullis [2] and later their potential was shown for on-demand drug delivery applications

[3]. However, preparation of protocells is still very complicated, therefore new methods of fabrication

need to be investigated before the possibility of industrial applications.

Practical part

Silica nanoparticles were prepared according to the literature [4] by the method firstly

published by Stöber [5] in the 30 - 60 nm range for intravenous applications [6] and the synthesis

proceeded as follows: 2.3 mL of Tetraethyl orthosilicate (TEOS) were mixed with a solution of 60 mL

of ethanol, 3.0 mL of ammonium hydroxide and 1 mL of water in a 100 mL flask being heated to 60 °C.

The reaction was left to proceed for 6 hours under 500 rpm stirring (stirrer length: 25 mm). The

prepared dispersion was dialyzed in ethanol for 2 days during which the ethanol was changed 5 times.

For better encapsulation into liposomes the nanoparticles were further aminated [7] with (3-

aminopropyl)triethoxysilane (APTES) using the following procedure: 1 mL of an APTES solution (3

mL ethanol, 50 µL deionized water, 63 µL APTES, adjusted to pH 2 with hydrochloric acid, mixed for

15 minutes prior use) was added into the dispersion containing 30 mL of silica nanoparticles (Approx.

10 mg/mL) and mixed in a room temperature for 24 h. The prepared dispersion was dialyzed in water

for 3 days during which the water was changed 10 times. Particles were stored in solution. The size

distribution of the prepared aminated silica nanoparticles was measured by dynamic light scattering

(DLS) and transmission electron microscopy (TEM). Results of this characterization is shown in Fig.

1. Both methods show diameter of particles around 50 nm.


Figure 1: Characterization of aminated silica nanoparticles: A) DLS size distribution; B) TEM image

In a typical synthesis of liposomes and protocells, 10 mg of mixture of lipids and cholesterol

were dissolved in 2 mL of mixture of chloroform : methanol (1 : 1) in a 25mL flask. Composition of

mixture of lipids and cholesterol was chosen due to literature [8] and previous experiences. After that

the solvent was evaporated in a rotary evaporator at 55 °C. The dried lipids formed a film around the

bottom of the flask. The sample was then kept under vacuum for at least 6 hours. For rehydration and

formation of liposomes, 2 mL of hydration dispersion containing water, NaCl, fluorescent agent

5(6)-carboxyfluorescein (5(6)-CF) and in case of protocells also silica nanoparticles dispersion was

added to the flask containing the dried lipid film and the flask was vortexed until the entire lipid film

was hydrated (no visible parts of the lipid film on the flask wall). Compositions of the hydration

solution for each sample is shown in Table 1. The sample was then transferred to a glass syringe

fitted in a liposome extrusion device heated to 69 °C and extruded through a porous membrane (pore

size 800 nm) 19 times to decrease the size of the formed liposomes or protocells.

Table 1: Protocells prepared for optimization of membrane and hydrating solution composition

Sample Liposome composition

DPPC : DPPG : Chol mol %

Hydrating solution

(always 2 mL) Fluorescent agent

1 85:0:15 6 mg/mL silica 5(6)-CF 7.5 mg/mL



4 45:40:15 6 mg/mL silica, 60

mmol/L NaCl

5(6)-CF 7.5 mg/mL


6 75:10:15 6 mg/mL silica, 60

mmol/L NaCl

5(6)-CF 7.5 mg/mL

Samples 1 and 2 aggregated in the process of hydration and were not studied further. Samples

3 – 6 were successfully prepared and their encapsulation efficiency was measured by a following

procedure: The sample was centrifuged (5000 RCF) and the supernatant was replaced with 60 mmol/L

water solution of NaCl. This was repeated three times. Then the samples were redispersed, left for 2

hours and then again centrifuged. Fluorescent emission spectra of the supernatant were measured, then

the samples were heated to 69 °C for 1 hour to promote release, centrifuged and another set of

fluorescent emission spectra were measured. Amount of the released fluorescent agent was calculated

from the difference of fluorescence intensity and is shown in Table 2.

A B


Table 2: Released amounts of 5(6)-CF from samples

Sample 5(6)-CF released (µg5(6)-CF)/mglipid)

1 x

2 x

3 6.3

4 11.3

5 5.6

6 16.2

Results show that the composition DPPC : DPPG : Chol of 75 : 10 : 15 is slightly better than

45 : 40 : 15 and this composition of lipidic membrane was used in further experiments. Because the

samples containing NaCl solution performed better, further experiments were made in the same

manner. Particle size distribution and TEM images were acquired for the sample 6. (Fig. 2)

Figure 2: Characterization of protocells - sample 6: A) DLS size distribution; B) TEM image

A set of fluorescent agents was chosen for further research in the field of protocells and

liposomes. To simulate various APIs fluorescein derivatives with similar structure but different

behaviour (solubility, polarity) were prepared: fluorescein (F), 5(6)-carboxyfluorescein (5(6)-CF),

fluorescein-5-isothiocyanate (FIC) and fluorescein-O-acrylate (FA). From all agents nearly saturated

solutions in water were prepared. Mass concentrations prepared solutions are shown in Table 3.

Table 3: Solutions used as hydration solution (always with NaCl)

Fluorescent agent Concentration / mg/mL

5(6)-carboxyfluorescein 7.5

fluorescein 2.5

fluorescein-5-isothiocyanate 0.3

fluorescein-O-acrylate 0

Fluorescein-O-acrylate proved to be so poorly soluble in water that further experiments with

this agent were not conducted. Protocells and liposomes loaded with the other three fluorescent agents

were prepared and the amount of the released fluorescent agents was calculated via the previously

described procedure. For liposomes the maximal theoretical capacity was calculated. Results are

shown in Table 4.

A B


Table 4: Released amount from liposomes and protocells for three different fluorescent agents

Sample Fluorescent agent

(mg/mL)

Hydrating solution

always 60 mmol/L

NaCl in 2 mL

Released amount of

fluorescent agent

(µgfluorescein)/mglipid)

Theoretical

capacity

(µgfluorescein)/mglipid)

7 5(6)-CF 7.5 Only solution 114.5 180

8 5(6)-CF 7.5 + 6 mg/mL of silica 16.2 x

9 F 2.5 Only solution 0 60

10 F 2.5 + 6 mg/mL of silica 0.8 x

11 FIC 0.3 Only solution 0 7

12 FIC 0.3 + 6 mg/mL of silica 25.2 x

The most soluble agent (5(6)-CF) (samples 7, 8) released 7 times more from pure liposomes

than from protocells and the released amount is comparable to the theoretical capacity. Fluorescein

(samples 9, 10) was not successfully encapsulated in liposomes and neither in protocells, although the

theoretical capacity is quite high. The release of the FIC (samples 11, 12) from liposomes was below

the detection limit, but was successfully released from protocells.

Conclusion

Silica nanoparticles were successfully encapsulated into liposomes with adsorbed

5(6)-carboxyfluorescein. The best composition of lipidic membrane was found for this molecule. It

was also shown that encapsulation and release provide better results in environment with NaCl (aq)

than in pure water. A set of protocells and liposomes loaded with three different fluorescent agents

(model APIs) was prepared and their release after heating was measured. From the three molecules

the Fluorescein-5-isothiocyanate provided better results when loaded into protocells compared to

liposomes where no release of Fluorescein-5-isothiocyanate was observed. This suggests that

protocells can be used as nanocarriers for APIs that cannot be encapsulated into pure liposomes and it

may lead into reconsideration of many APIs that were in history rejected for liposomal formulation

due to complications with their encapsulation.

References

1. Allen, T.M. and P.R. Cullis, Liposomal drug delivery systems: from concept to clinical

applications. Advanced Drug Delivery Reviews, 2013. 65(1): p. 36-48.

2. Ashley, C.E., et al., The targeted delivery of multicomponent cargos to cancer cells by

nanoporous particle-supported lipid bilayers. Nature Materials, 2011. 10(5): p. 389-97.

3. Butler, K.S., et al., Protocells: Modular Mesoporous Silica Nanoparticle-Supported Lipid

Bilayers for Drug Delivery. Small, 2016. 12(16): p. 2173-85.

4. Oh, W.-K., et al., Cellular uptake, cytotoxicity, and innate immune response of silica-Titania

hollow nanoparticles based on size and surface functionality. ACS Nano, 2010. 4(9): p. 5301-

5313.

5. Stöber, W., A. Fink, and E. Bohn, Controlled growth of monodisperse silica spheres in the

micron size range. Journal of Colloid and Interface Science, 1968. 26(1): p. 62-69.

6. Tarn, D., et al., Mesoporous silica nanoparticle nanocarriers: Biofunctionality and

biocompatibility. Accounts of Chemical Research, 2013. 46(3): p. 792-801.

7. Vinklář, O., Investigation of protein-nanoparticle affinity. 2016, UCT Prague.

8. Haša, J., J. Hanuš, and F. Štěpánek, Magnetically controlled liposome aggregates for on-

demand release of reactive payloads. ACS Applied Materials & Interfaces, 2018. 10(24): p.

20306-20314.


1

Preparation of protocells: encapsulation efficiency and on-demand payload release

Filip Hládek

Supervisor: Prof. Ing. František Štěpánek, Ph.D.

Supervisor-specialist: Ing. Denisa Lizoňová

Keywords: protocells, mesoporous silica nanoparticles, liposomes, controlled release, drug

carrier

1. Introduction

The majority of newly synthesized drugs is characterized by poor water solubility, which can be

improved either by amorphization or by loading into a carrier. The goal of this work is to prepare

so-called protocells - mesoporous silica nanoparticles encapsulated in liposomes. The silica

nanoparticle acts as a carrier for a drug, while the liposome acts as a gatekeeper which prevents

any undesired drug leakage.

The properties of the mesoporous silica nanoparticles below 200 nm are very suiting for this

application. High surface to volume ratio (thanks to the mesoporosity) allows good load capacity

while remaining small enough to be viable option for a medical use (possible intravenous

application).

The liposome capsule is a spherical layer consisting of at least one lipid bilayer that prevents

leakage of its contents, in this case the drug. The drug release can be done by different

approaches aiming to the liposome layer disruption, for example by heating the protocell [1].

Even thou there is a minor release observed at room temperature, full cargo release is usually

achieved when exceeding 70°C.

The protocells can be applied in intravenous, oral or dermal medical application, depending on

their size and the desired effect. One of the proposed applications is to transport the protocells to

the desired compartment of the body, where the drug effect is needed, and release it there. For

the release in the human body the different approaches have to be considered, for example the

disruption of the liposome by the radiofrequency heating in the presence of super-paramagnetic

nanoparticles, which is currently studied in our research group.

2. Experiments

2.1. Silica nanoparticles synthesis and optimization

2.2. Synthesis

Synthesis of the mesoporous silica nanoparticles was based on SolGel approach [2]. Silica

nanoparticles after the synthesis and re-suspendation (ultrasonic needle) are shown in Figure 1.


2

Figure 1. Mesoporous silica nanoparticles.

2.3. Optimization

There was a need for adjustments due to the inconsistency of CTAB effect on synthesis.

According to the Zhaogang [2] by increasing the CTAB concentration from 6 mM to

15 mM, size of the nanoparticles decreases by approximately 40%. Unfortunately, after the

initial experiments the original flask of CTAB was depleted and it was found out that the

synthesis described in [2] does not work properly with different batches of CTAB, thus

optimization was necessary. That could be done by either changing water to ethanol ration

during synthesis or by increasing CTAB concentration.

After the thorough examination, the best solution for that specific batch of CTAB was to keep

the EtOh: H2O ratio while increasing the CTAB concentration to 30mM and keeping the

synthesis at 30°C. TEM pictures of silica nanoparticles before and after optimization are shown

in Figure 2.

Figure 2. Silica particles before (left) and after (right) optimization.


3

3. Characterization

3.1. TEM

The Transmission Electron Microscopy (TEM) pictures were used to confirm the particle

mesoporosity and size, and to confirm that the size measurement by Malvern Zetasizer Nano-ZS

is accurate enough. All further measurements of the particle size distribution were done by

Zetasizer Nano (Dynamic Light Scattering).

3.2. Size

In order to keep all experiments under relatively constant conditions, the size of the mesoporous

silica nanoparticles was required to be approximately 180 ± 20 nm. To make sure all synthetized

particles were in this range, a controlling measurement was made after every synthesis and

before each step of the loading and encapsulation. Size distributions during various stages of

preparations are shown in Figure 3.

Figure 3. Volume distribution after synthesis (orange), after ultrasonic dispergation in distilled water

(gray), after loading with 5-(6)-Carboxyfluorescein (green) – described in the following chapter and after

encapsulation without extrusion (blue).

4. Particle loading

In order to determine the load capacity of the silica nanoparticles, two batches of particles were

prepared. One was loaded 1:1 by weight with 5-(6)-Carboxyfluorescein (later 5(6)-CF), the other

in 1:2 ratio. Both batches were measured by Fluorescence Spectrophotometer. Based on the

measured fluorescence, the residual concentration of 5(6)-CF in solution was calculated as well

as the amount of 5(6)-CF that was successfully loaded into the particles. Results of the particle

loading are shown in Table 1.

Table 1. Particle loading. Absorbance of calibration solution and samples concentrations based on

calibration curve.

Sample Initial concentration

(mg/ml)

Fluorescence

(a.u.)

Residual concentration

after loading (mg/ml)

Load capacity

(mg5(6)-CF/mgsilica)

Calib. 1,9 237 1,91 x

1:1 ratio 5 592 4,94 0,006

1:2 ratio 5 413 3,43 0,0785

0

5

10

15

20

25

30

35

40

45

0 200 400 600 800 1000 1200 1400 1600 1800 2000

Vo

lum

e %

Size (nm)


4

5. Protocells preparation

Protocells were prepared according to protocol described by Balouch, 2017 [3], originally using

non-porous silica nanoparticles with 50 nm diameter. Liposome film used consisted of

dipalmitoylphosphatidylcholine (DPPC, 6.5 mg), 1,2-Dipalmitoyl-sn-glycero-3-

phosphorylglycerol (DPPG, 2 mg) and cholesterol (2.5 mg). The extrusion with mesoporous

silica nanoparticles was unsuccessfully attempted on 800 nm membrane, as well as 1000 nm

membrane. Since all attempts to replicate Bc. Balouch’s experiments with the larger and

physically different silica nanoparticles were unsuccessful yet, different approaches have to be

studied.

5.1. Further experiments

Since previous attempts to encapsulate a single nanosilica particle into the liposome shell were

unsuccessful, other experiments are planned. One possible way would be to dilute the hydration

solution to decrease the chance that more than one particle will be encapsulated. Another

possibility would be not to hydrate the phospholipid film directly, but dissolve it in distilled

water first, then add part of the silica suspension and then extrude. This process would be

repeated until about 2 ml of the mesoporous silica nanoparticle suspension is used.

6. Release

Initial plans included release experiments from the protocells, but due to the problems with the

synthesis optimization and protocells preparation/extrusion, these experiments were not carried

out yet. The plan is to centrifuge the suspension, remove the supernatant and replenish the vial

with solution with correct osmolarity, then measure the supernatant absorbance, temper the

solution above 80°C to destroy the liposomes and release the 5(6)-CF. Then measure the

absorbance again and calculate the concentrations based on the calibration curve.

7. Conclusion

Results of these experiments seem promising. The protocells prepared by the hydration of the

phospholipid layer using mesoporous silica nanoparticles suspension resulted in the objects too

large for the extrusion using 1000nm membrane. However, such protocells can be used in oral or

dermal applications. More experiments regarding encapsulation and release kinetics need to be

carried out to in order to determine the real potential of the protocells consisting of the

mesoporous silica nanoparticles and the liposome gatekeeper.

8. Acknowledgements

The author would like to thank Bc. Martin Balouch for Transmission Electron Microscopy

measurements of prepared nanoparticles.

9. References

[1] M. Ulrich, J. Hanuš, J. Dohnal, F. Štěpánek, Journal of Colloid and Interface Science,

394 (2013), 380-385

[2] T. Zhaogang, T. Yandong, H. Jun, L. Feng, Y. Wensheng, Y, Microporous and Mesoporous

Materials, 127 (2010), 60-72


5

[3] M. Balouch, Fabrication of nano-sorbents and their encapsulation into liposomes, Bachelor

thesis, 2017


Light-shaped chitosan hydrogels

Kristýna Idžakovičová

školitel: RNDr. Ivan Řehoř, Ph.D.

Keywords: hydrogels, stop-flow lithography, chitosan, methacrylated dextran

1) Current knowledge

Hydrogels are cross-linked materials capable of absorbing large amounts of water without

dissolving. Due to their high water content and soft consistency, they closely resemble natural living

tissue. Therefore, there have risen a need to develop methods to create hydrogel microparticles that

would qualitatively simulate individual cells in size and functionality. So-called microgels can serve

a great purpose in many bioapplications – as carriers for drug delivery [1] or as artificial cells [2].

For all those tasks, microgels are required, having specific shape, size, material composition and

other characteristics. Ideal synthetic method of microgels would allow creating highly uniform

hydrogels that have well-defined shape. Furthermore, this method would also allow to load

microgels with cargo of choice, ranging from molecules to microparticles and living cells. Because

classic bottom-up processes do not provide control over microgel shape and compartmentization,

lithographic top-down method was chosen for this work. Many lithographic methods are well-known

nowadays and also successfully work in fields of bioapplications [3] , namely soft-lithography [4],

photolithography [5] and stop-flow lithography [6] (SFL). General problem of these methods is a

low production rate. Except of SFL where production rate goes up to million particles per hour.

Another drawback represents the need to have photocrosslinkable polymers for lithographic

processing [3]. Implanting photocrosslinkable groups into biopolymers can severely change their

original properties, mainly biocompability which can be very non-beneficial. Synthetic polymers are

already widely used in biomedical research, they are xenogenous material therefore they can have

harmful, undesired effect on human body [7]. To avoid these undesired behaviours, it is better to use

solely biopolymers in synthesis of microgels. Biopolymers can be crosslinked using various

methods, but light is not one of them, which prevents their lithographic processing. Thus, it is

required for current bioapplications to find a synthetic method that can produce shaped microgels

with high production rate from biopolymers.

2) Principle of the work

The synthetic method is depicted in the Figure 1 and consists of following steps. First the degradable

photocrosslinkable polymer dex-HEMA (blue) is mixed with biopolymer (chitosan - red). Dex-

HEMA consists of natural polymer – dextran modified with 2-hydroxyethylmethacrylate moieties

along the chain. Because the connection between dextran and HEMA is achieved through

hydrolytically labile carbonate esters, hydrogels prepared by crosslinking dex-HEMA degrade under

physiological conditions [8]. Microgels of required shape are created from this polymer using stop-

flow lithography. After collecting and washing the synthetized microgels, the biopolymer is

crosslinked, forming a network with dex-HEMA. Genipin is used as crosslinking agent for

biopolymer. Then, the dex-HEMA is hydrolyzed and washed away under basic conditions yielding

solely biopolymer particles.


Figure 1: Scheme of proposed synthetic method

First, polymer blend from photocrosslinkable polymer (Dex-HEMA - blue) and biopolymer

(chitosan - red) is prepared. Microgels are synthetized from the blend by local crosslinking of

photocrosslinkable Dex-HEMA with Stop-Flow lithography. Chitosan inside formed microgels is

crosslinked using genipin. Then the Dex-HEMA is hydrolyzed and washed away, leaving pure

chitosan anisotropic microgels.

Figure 2: Scheme of the SFL microgel synthesis

3) Synthesis of particles

Templating polymer (Dex-HEMA) was mixed with chitosan and photoinitiator. The

conditions were found to provide single phase mixtures (e.g. higher photoinitiator concentration

induce phase separation), of short photocrosslinking time, low viscosity and high chitosan content.

The composition of the mixture, fulfilling these often antagonistic requirements was: 15 % of dex-

HEMA, 2,5 % of chitosan and 0,25 % of photoinitiator. Higher chitosan concentration would result

in high viscosity which would significantly aggravate processing in microfluidic channel.

To enable visualization of chitosan with fluorescence microscopy, we labelled it with a

fluorescent dye following this protocol: We prepared 100 ml solution by dissolving 1 g of chitosan in

0,25% acetic acid. Then we added 5 mg fluoresceinisothiokyanate (fitc) into solution and stirred the

solution for 2 hours hidden from light in laboratory temperature. We then dialysed fluorescein out of

the solution. Lastly, we used rotary evaporator to thicken solution gaining approximately 4 %

solution of fitc-labeled chitosan. We prepared 4 % solution of non-labeled chitosan (4 % chitosan and

1 % acetic acid) and added 1 ml of fitc-labeled chitosan. This procedure ensured precise

concentration of chitosan. The stock solution was then prepared by diluting previously mentioned

solution with 1 % acetic acid.


After we prepared solution of dex-HEMA and chitosan of required concentrations, we added

2 mg of rhodamin to label dex-HEMA and 5 mg of 5 % stock solution of photoinitiator (LAP).

Solution was then centrifuged, and sample was prepared for processing in microfluidic channel. The

principle of stop-flow lithography is schematically shown in Figure 2. The liquid mixture is pumped

through microfluidic channel. The flow is then stopped, and the area defined by photomask is

irradiated with UV light. Crosslinking reaction occurs, and solid particle is formed. The flow is then

reestablished, and particle is washed away, and the process is repeated. Higher viscosity liquids,

respectively microgels formed of them tend to stick to the walls of the channel and block effective

flushing of the channel, hampering the entire process, therefore we were periodically cycling several

irradiation positions inside the channel, using a programmed custom built motorized stage, coupled

to the SFL control unit. Particles were then collected and washed with water repeatedly. Figure 3A

shows particles before crosslinking with genipin. 10 mM of genipin solution was prepared by

dissolving genipin in 1 % acetic acid and added into wells with particles. Incubation time of

crosslink of chitosan was one and half hour at 37 °C. Particles turned blue which indicated successful

crosslinking reaction [9]. After removing genipin solution and washing particles with water, we

added 1 M solution of NaOH to hydrolyze dex-HEMA. This process was not as effective as we

initially expected, and we had to cycle water/ NaOH 3 times leaving approximately 5 minutes

between each change. After hydrolysis of dex-HEMA, we also added 2 % acetic acid to dissolve any

non-crosslinked chitosan and to see if they are dissolving like they did in only-dex-HEMA-

crosslinked state. Then we again added 1M solution of NaOH and left them for 5 days. These

particles are shown in Figure 3B (in bright field of light) and Figure 3C (in fluorescent field of light).

Figure 4 shows comparison of decomposing dex-HEMA if chitosan is crosslinked (4A) and is not

crosslinked (4B).

Figure 3: Pictures of synthetized microgels (description of pictures is written in the text; all

scales are 100 μm)

Figure 4: Particles after addition of 1 M NaOH (A) chitosan crosslinked, (B) chitosan not

crosslinked (both scales are 100 μm)

4) Results and discussion

We successfully prepared mixture of biopolymer (chitosan) and template polymer

(methacrylated dextran). Synthetic polymer was crosslinked via stop-flow lithography to gain desired

A C B

A B


shape and then we used genipin to crosslink entrapped biopolymer. Synthetic polymer was then

hydrolysed under basic conditions yielding mainly biopolymer particles.

Major problem we encountered is ability of chitosan to effectively hamper hydrolysis of dex-

HEMA. Chitosan is pH-responsive polymer, which can only be dissolved under acidic conditions.

Adding NaOH (basic) into solution results in precipitation of chitosan (observed as a shrinkage of

the hydrogel) which prevents desired hydrolysis. One idea how to solve this, is to use more time

between washing cycles of water and NaOH, giving chitosan time to loosen in neutral conditions.

After this goal is accomplished, we want to find conditions for self-assembly of biopolymer particles.

We also plan to use phytic acid as an ionic crosslinking agent for chitosan.

5) Acknowledgement

I would like to thank my mentor RNDr. Ivan Řehoř, Ph.D. for his great help and guidance in

this project.

6) References

[1] S. Mitragotri, P. A. Burke, R. Langer, Nat. Rev. Drug Discov. 2014, 13, 655.

[2] T. J. Merkel, S. W. Jones, K. P. Herlihy, F. R. Kersey, A. R. Shields, M. Napier, J. C. Luft, H. Wu,

W. C. Zamboni, A. Z. Wang, J. E. Bear, J. M. DeSimone, Proc. Natl. Acad. Sci. 2011, 108, 586.

[3] M. E. Helgeson, S. C. Chapin, P. S. Doyle, Curr. Opin. Colloid Interface Sci. 2011, 16, 106.

[4] T. T. Dang, Q. Xu, K. M. Bratlie, E. S. O’Sullivan, X. Y. Chen, R. Langer, D. G. Anderson,

Biomaterials 2009, 30, 6896.

[5] J. L. Perry, K. P. Herlihy, M. E. Napier, J. M. DeSimone, Acc. Chem. Res. 2011, 44, 990.

[6] D. Dendukuri, S. S. Gu, D. C. Pregibon, T. A. Hatton, P. S. Doyle, Lab. Chip 2007, 7, 818.

[7] A. O. Elzoghby, W. M. Samy, N. A. Elgindy, J. Controlled Release 2012, 161, 38.

[8] W. N. E. van Dijk-Wolthuis, J. A. M. Hoogeboom, M. J. Van Steenbergen, S. K. Y. Tsang, W.

E. Hennink, Macromolecules 1997, 30, 4639.

[9] Butler, M. F.; Ng, YF;, Y.; Pudney, P. D. A. Chin. J. Polym. Sci. 2003, 41, 24


The Effect of pH of Supply Solutions on pH Dynamic Behaviour of Urea-Urease

System in CSTR

Emel Ilgın Karakoç

Supervisor: Ing. František Muzika, Ph.D.

Keywords: pH oscillations, Urea-Urease, enzymatic reactor

Oscillations causes increase or decrease of concentration of some chemicals in the system, it

makes the system dynamic, which means, the system is away from thermodynamic equilibria. It

could be naturally occurring oscillations, that are observable in plants or during glycolysis in the

yeast cells. Oscillations occur as a synergic effect of autocatalysis (positive feedback) and inhibition

(negative feedback), where both can have different forms. Oscillations in urea-urease system were

proposed in a model of lipid vesicles, where the driving force of oscillation is unequal ratio of

transport rate coefficient of urea and H+ ions [1], which helps the reaction to provide negative

feedback (depletion of substrate) and build-up of H+ ions. In this study we used sulfuric acid as

primary source of H+ ions[2]. Urea hydrolysis with urease enzyme creates weak base, that contains

ammonia and carbon dioxide, which in water gives bicarbonate. Urea, urease and acid follows

reaction series that is shown in Eq. 1. Urea-urease reaction occurs in nature in Helicobacter Pylori,

which is trying to protect itself from acid in stomach by increasing pH, causing gastric ulcer. These

reactions partially follows the Michaelis-Menten and they have bell shaped activity curve from pH 3

to 11 showing maximum at pH 7 [1].

2 2

3 3

4 4

5 5

2 2 2 3 2

4 3,

2 2 3,

2

3 3,

2,

( ) 2

r

r

r

r

urease

k k

k k

k k

k k

CO NH H O NH CO

NH NH H

CO H O H HCO

HCO CO H

H OH H O

(1)

ir

i

k reverse equilibrium constant

k forward equilibrium constant

i = number of the reaction

The urease enzyme occurs in the nature in different forms and from different sources. In our

study we used urease from Jack beans (Canavalia ensiformis). The purpose of production of urease

in Jack bean is to protect legume from insects. Other than plants, bacteria and fungi can also create

urease [3].

Urea-Urease system is promising target for the bio regeneration of materials. The cracks in

concrete can be filled with solution of powdered chalk (CaCO3) dissolved in lactic acid creating a

Ca2+ solution and then other solution of urea and urease should be added. The urease solution

hydrolyze urea producing ammonia and CO2, which dissolves in the water forming bicarbonate ions,

that can precipitate and produce calcium carbonate [4]. Recently studies showed us, that breathing

microgel can use acid solution as ‘fuel’ to keep system in dynamic state, in this way we can keep the


system in and out of equilibrium by using pH changes [5]. Increasing of pH affect the activity of

urease.

In contrast to classical drug delivery methods researchers found an alternative way to deliver

drugs to the certain locations by using self-propel aqueous solutions. However, this phase is

generally based on toxic fuels such as hydrogen peroxide and hydrazine. Recent studies show us we

can use the same way, but more biocompatible. They are covering the silica gels with urease enzyme,

which can create mechanical energy in ionic media. Advantages of using urease nanomotors are high

loading capacity and active transport toward to the cancer cells [6].

In this study we used CSTR with three inflows and volume of the reactor was 2.86ml. Our

first inflow was 0.0015M stock urea with F0~0.08ml/min. Second inflow was stock urease with

5U/ml, and F0~0.33ml/min and the third flow was stock sulfuric acid solutions with pH=<2.624;

4.81>, and F0~0.08ml/min. Experiments took place under 25°C, controlled by A/C. Different stock

concentrations of acid were used to find dynamic behaviour. In the Fig. 1 we can see the schematic

representation of the experimental setup. Before and after measurement the system was cleaned with

chlorine and water. In the first step the pH meter (11) and the pump (5) were calibrated. pH was

measured using pH probe theta 113vfr (2) and pH meter HI 5222-02 (11). After that we started to

feed our system with stock solutions (6, 7, 8) with high flow rate for 5 minutes and then flow rate

was decreased to calibrated value and pH started to be recorded, later it was analyzed from graphical

representations of pH recordings over time.

Figure 1. The scheme of experiment (1) Reactor, (2) pH probe theta 113vfr, (3) Stirrer bar, (4)

Magnetic stirrer Wisd MSH-20D, (5) Pump ISMATEC IPC-N8 , (6) Urea solution, (7) Urease

solution, (8) Sulfuric acid solution, (9) Waste tank, (10) Thermometer probe, (11) pH meter HI5522-

02

In the Fig. 2 some of our experimental results can be seen. We can observe different behaviour

under different sulfuric acid concentration levels. Sometimes urea-urease reacted and stayed

approximately in the same pH level (equilibrium), in some cases we observed unreacted states (no

apparent increase of pH due to formation of bicarbonates and ammonia) and pH oscillations. As we

can see oscillations occurred between pH 5 and 5.5 in red and yellow curves and between pH 6.7 and

7.0 for violet curve.


Figure 3. Experimental recording of pH in time for various H2SO4 concentrations and residence time

300 seconds

Figure 3. Experimental recording pf pH for residence time 32 seconds

As we can see in the Fig. 3 oscillations occurred between pH 4 and 6.5 in the curve. We can

see sharp decrease and increase of the pH in the system. In this particular experiment, we used higher

flow rate, therefore residence time was decreased to 32 seconds. We can observe smaller pH

oscillations from the start, but at 8000, 13000 and 15000 seconds, the pH oscillations were ranging

from ΔpH~1.5 to ΔpH~2.0. After 16000 seconds the pH of the system was slightly fluctuating

around value 6.5.


Our results could provide periodic behavior for breathing hydrogels[5], rubber muscle

actuation[7], nano-minerals preparation and ground water treatment[8], bio-cementation[4] and

water treatment[9].

References

[1] Miele, Y., et al., Modelling Approach to Enzymatic pH Oscillators in Giant Lipid Vesicles, in

Advances in Bionanomaterials. 2018. p. 63-74.

[2] Bubanja, I.N., T. Bánsági, and A.F. Taylor, Kinetics of the urea–urease clock reaction with

urease immobilized in hydrogel beads. Reaction Kinetics, Mechanisms and Catalysis, 2017.

123(1): p. 177-185.

[3] Carrazoni, T., et al., Jack bean urease modulates neurotransmitter release at insect

neuromuscular junctions. Pestic Biochem Physiol, 2018. 146: p. 63-70.

[4] Phua, Y.J. and A. Røyne, Bio-cementation through controlled dissolution and

recrystallization of calcium carbonate. Construction and Building Materials, 2018. 167: p.

657-668.

[5] Che, H., S. Cao, and J.C.M. van Hest, Feedback-Induced Temporal Control of "Breathing"

Polymersomes To Create Self-Adaptive Nanoreactors. J Am Chem Soc, 2018. 140(16): p.

5356-5359.

[6] Hortelão, A.C., et al., Enzyme-Powered Nanobots Enhance Anticancer Drug Delivery.

Advanced Functional Materials, 2018. 28(25).

[7] Sutter, T.M., et al., Rubber muscle actuation with pressurized CO2from enzyme-catalyzed

urea hydrolysis. Smart Materials and Structures, 2013. 22(9).

[8] Abdel-Gawwad, H.A., S.A. Mohamed, and A.H. Mostafa, Application of eco-friendly method

for nano-minerals preparation and ground water treatment. Ecological Engineering, 2018.

119: p. 29-34.

[9] Veaudor, T., et al., Overproduction of the cyanobacterial hydrogenase and selection of a

mutant thriving on urea, as a possible step towards the future production of hydrogen coupled

with water treatment. PLoS One, 2018. 13(6): p. e0198836.


SYNTHESIS AND CHARACTERIZATION OF NANOPARTICLES USING

MICROFLUIDICS

Aliye Hazal Koyuncu

Supervisor: Ing. Viola Tokárová PhD

Keywords: nanoparticles, microfluidics, silver, flow-focusing

1. Introduction

For the nanoparticle-based applications, controlling the particle size and producing a unique size and

shape of particles are important due to their physico-chemical properties. Traditionally batch

synthesis methods are the most common ones for nanoparticle production [1]. However, in the batch

methods, it is often hard to control bulk shearing forces and the methods are not flexible or efficient

to obtained reproducible particles with monodisperse or narrow size distribution [2]. Therefore, there

is an increasing demand to control reaction parameter for nanoparticle synthesis [3]. Microfluidics is

a promising method for nanoparticle synthesis overcoming limits of batch synthesis methods [4].

Firstly, a large surface area to volume ratio of microchannels helps to increase mass and heat transfer

in the system, which increases the efficiency of the reaction in a smaller volume compared to a high-

volume batch methods. Secondly, microchips are more suitable to work in harsh conditions in

comparison to the batch reactors with regards to rapid temperature or pressure changes while using

toxic and/or explosive materials [5].

2. Experimental Part

2.1. Material

Silver nitrate (AgNO3, Sigma, purity > 99 %), trisodium citrate dihydrate (C6H5Na3O7.2H2O, Penta,

purity > 99 %), tannic acid (C76H52O46, Sigma Aldrich), polydimethylsiloxane (PDMS, Sylgard 184,

Dow Corning), silicone elastomer (Sylgard 184, Dow Corning), mineral oil (Sigma), diethyl ether

(C4H10O, Penta), demineralized water (Aqual 25, conductivity ~ 0.07 μS/cm).

2.2. Batch Method

The synthesis of silver nanoparticles in the batch was done by the precipitation of silver nitrate as the

silver source according to Ranoszek-Soliwoda et al. [6]. An aqueous solution of silver nitrate

(0.9625 mM, 95 mL), trisodium citrate dihydrate (0.07 M, 4.03 mL) and tannic acid (0.0015 M,

0.60 mL) were prepared using demineralized water. The molar ratio of the silver nitrate, trisodium

citrate dihydrate, tannic acid was calculated to be 1.0:3.1:0.1. Silver nitrate solution

was placed in a 250 ml glass round bottom flask, and the mixture of silver nitrate solution with the

tannic acid solution was added dropwise into the flask at room temperature under continuous stirring

at 500 rpm. When the addition was completed, stirring was kept for 15 more minutes. The samples

were analysed by a dynamic light scattering (DLS) and UV-VIS spectrophotometer. Additionally,

nanoparticles were visually characterized by a transmission electron microscopy (TEM).


2.3. Microfluidic Method

2.3.1. Microfluidic Chip Design and Preparation

The microchips were prepared by a soft lithography method using polydimethylsiloxane (PDMS)

and a silicon wafer serving as a master mould. The PDMS was prepared by mixing the curing agent

and the silicone elastomer (1:5 w:w) and poured over the silicon wafer placed in a petri dish. The

polymer mixture was degassed and heat-treated at 75 °C for 22 minutes. The crosslinked PDMS was

peeled off from the wafer and holes were punched for the channel inlets

and outlet. The bottom part of the chip

was prepared using a PDMS

mixture of 1:14 w:w ratio (curing

agent to silicone elastomer) poured

inside an empty petri dish,

degassed and put in the oven at 75 °C

for 20 minutes. The surface of the

bottom PDMS layer remained sticky,

and the upper part of the chip with

channels was placed on top of it.

2.3.2. Microfluidic Process

Silver nanoparticles were synthesized by the reduction reaction of silver nitrate using trisodium

citrate dihydrate solution as a reduction agent. Briefly, silver nitrate (SN, AgNO3) solution (1 mL

and 0.92 mM), trisodium citrate dihydrate (SC, C6H5Na3O7.2H2O) solution (1 mL and 5.72 mM) and

tannic acid (TA, C76H52O46) solution (1 mL and 0.185 mM) were prepared with demineralized

water. The molar ratio of all reactants (1.0:3.1:0.1) was kept the same as in the batch process. The

both disperse phases - SN and SC with TA solutions were pre-filtered before use. SN solution (1 ml)

was transferred into a 1 ml volume glass syringe (Hamilton). Then, 1 ml of SC and TA solution was

transferred to the other glass syringe (Hamilton) of the same volume. Finally, 2 ml of a mineral oil

(continuous phase) was transferred into a 2.5 ml volume glass syringe. All syringes were connected

to the microchip via PTFE capillary tubing. The flow rates were set for each phase and experimental

setup individually by the linear pump software (neMESYS). The final product was collected into

sample tubes pre-filled with 0.5 mL of demineralized water to quench the reaction and particle

growth. Then the W/O emulsion was separated by centrifugation for 3 minutes at 13.4 rpm

(Eppendorf, MiniSpin) to oil and aqueous phases. After discarding of the oil phase, samples were

further cleaned by the extraction with diethyl ether to remove remaining traces of the oil.

Figure 1 Microfluidic chip: A) silicon wafer with the chip

design including two disperse phases (1) and (2) and one for

the continuous phase (3), the flow-focusing junction (4),

where the droplet generation occurs, and one output (5) ; B)

the detailed image of the flow focusing part.


Two different methods were

used to produce silver

nanoparticles using a

consistent molar ratio of

reagents (1.0:3.1:0.1) in the

system as described in the

Experimental Section. The

dynamic behaviour of the

batch synthesis was examined using dynamic light scattering (DLS) and no linear growth of particles

in time was observed as expected. Additionally, silver nanoparticles have polydisperse distribution.

On the other hand, particles prepared in the microfluidic chip show less polydisperse populations

than in the case of batch production (Table 1 Experimental conditions (concentration and flow rates) of

all reagents for silver nanoparticle synthesis

Sample

CSN

[mM]

Csc

[mM]

CTA

[mM]

QSN

[µl/h]

QSC and TA mixture

[µl/h]

QSC TA+QSN

[µl/h]

Qmineral oil

[µl/h]

1 0.92 5.72 0.185 8 8 16 60

2 0.92 5.72 0.185 8 8 16 80

3 0.92 5.72 0.185 16 16 32 60

4 0.92 5.72 0.185 16 16 32 80

A). Only one sample with the set parameters (sample 3 from Table 1) shows polydispersity caused

by the presence of many nuclei of the silver nanoparticles with a size of few nanometres. The similar

trend was observed by evaluating the TEM images (Table 2) where Sample 3 has the largest

standard deviation of all prepared samples. All DLS data are shown in the result Table 2.

Additionally, the UV-VIS spectra were measured for all samples, proving the obtained samples are

silver nanoparticles with λmax = 420 nm.

Table 1 Experimental conditions

(concentration and flow rates) of

all reagents for silver

nanoparticle synthesis

Sample

CSN

[mM]

Csc

[mM]

CTA

[mM]

QSN

[µl/h]

QSC and TA mixture

[µl/h]

QSC TA+QSN

[µl/h]

Qmineral oil

[µl/h]

1 0.92 5.72 0.185 8 8 16 60

2 0.92 5.72 0.185 8 8 16 80

3 0.92 5.72 0.185 16 16 32 60

4 0.92 5.72 0.185 16 16 32 80

The average sizes and their standard deviations for particles prepared using different flow rates of

dispersed and continuous phases are shown in the

Table 2. The results of DLS are defined based on the volumetric size distribution. The results show

that increased flow rate of the dispersed phase from 16 to 32 µl/h and fixed continuous phase lead to

Figure 2 (A) Volume size distribution graphs of Ag nanoparticle

samples produced by microfluidics with various flow rates; (B) UV-

Vis spectra of Ag NPs prepared by both methods.


increased values of the final particle size measured by both, DLS (27.7 nm to 46.1 nm) and the

image analysis of TEM images (12.8 nm to 22.3 nm). The results of the batch process show the

biggest size values of the silver nanoparticles.

Table 2 Influence of the flow rates of dispersed and continuous phases on particle size distribution for

microfluidic experiments and comparison with the batch process

Sample

Qdisperse

[µl/h]

Qcontinuous

[µl/h]

Particle Size

DLS (nm)

Particle Size

TEM (nm)

1 16 60 32.8 ± 7.9 16.3 ± 9.0

2 16 80 27.7 ± 8.3 12.8 ± 9.9

3 32 60 35.7 ± 5.6 20.7 ± 14.4

4 32 80 46.1 ± 9.1 22.3 ± 8.2

Batch - - 65.2 ± 26.5 42.4 ± 29.1

TEM images show the final particles obtained from the microfluidic process are mostly spherical and

relatively monodisperse (Figure 3). Also, no agglomeration was observed. The batch product shows

more polydisperse and larger particles having more irregular shape compared to microfluidic

products.

Figure 3 TEM images of final silver nanoparticles from microfluidic (A-D) and batch (E) process: A) sample

1 B) sample 2, C) sample 3, D) sample 4 (all microfluidic samples correspond with the designation in

Table 1), E) sample of the batch process

3. Conclusion

In this work, the experimental results presented that both methods, standard batch and microfluidic,

were successful in silver nanoparticle production. However, in comparison to the batch process,

particle obtained from the microfluidic process have narrower size distribution and more spherical

morphology. Additionally, due to the susceptibility of silver nanoparticles to light absorption

(yellow-brown colour change), the particle size and shape could be measured during the process and

their final properties could be controlled.

Acknowledgements

This work has been supported by the Czech Science Foundation (project GACR 17-11851Y).

References

[1] S. S. Eun, Y. Sunhye, C. H. Hee, and C. Soonja, “Fully crosslinked poly(styrene-co-divinylbenzene)

microspheres by precipitation polymerization and their superior thermal properties,” J. Polym. Sci.

https://obd.vscht.cz/fcgi/verso.fpl?fname=g_grant_sez&_uloz_filtr=ano&_cislo=17-11851Y


Part A Polym. Chem., vol. 42, no. 4, pp. 835–845.

[2] G. A. Patil, M. L. Bari, B. A. Bhanvase, V. Ganvir, S. Mishra,“Continuous synthesis of functional

silver nanoparticles using microreactor: Effect of surfactant and process parameters,” Chem. Eng.

Process. Process Intensif. pp. 69–77, Dec. 2012.

[3] J. Wang et al., “Droplet Microfluidics for the Production of Microparticles and Nanoparticles,”

Micromachines, vol. 8, no. 1, p. 22, Jan. 2017.

[4] C.-X. Zhao, L. He, S. Z. Qiao, and A. P. J. Middelberg, “Nanoparticle synthesis in microreactors,”

Chem. Eng. Sci., vol. 66, no. 7, pp. 1463–1479, 2011.

[5] N. Hao, Y. Nie, and J. X. J. Zhang, “Microfluidic synthesis of functional inorganic micro-

/nanoparticles and applications in biomedical engineering,” Int. Mater. Rev., pp. 1–27, 2018.

[6] K. Ranoszek-Soliwoda et al., “The role of tannic acid and sodium citrate in the synthesis of silver

nanoparticles,” J. Nanoparticle Res., vol. 19, no. 8, p. 273, Aug. 2017.

SVK 2018, UCT Prague, Department of Chemical Engineering

Design of Dissolution Method for Poorly Soluble Drug Formulation

Bc. Erik Sonntag

supervisor: prof. Ing. František Štěpánek, Ph.D.

INTRODUCTION

Sustained delivery of a drug at a target site is an important theme in controlled drug-delivery

systems. Commonly used drug-delivery systems capable of releasing the drug for more than a

week are parenteral injections. These long-acting injectable formulations offer many advantages

including: a predictable drug-release profile; improved systemic availability by avoidance of

first-pass metabolism; decreased incidence of side effects; better patient compliance and overall

cost reduction of a medical care [1].

Sustained release injectable suspensions are one type of the long-acting formulations. Their unique

feature is the creation of a depot at the injection site after the drug administration [2]. This work

deals with the study of the specific intramuscular suspension of this type. Solid particles in this

formulation are present in a form of a poorly water soluble prodrug which is a compound that,

after administration, is metabolized into an active pharmaceutical ingreedient (API). This process

is called an enzymatic dissolution and can be divided into two parts. The first is a physical

dissolution of solid particles into the tissue fluid. Once the solid particles enter a liquid phase,

enzymatic conversion of prodrug to the API occures. Particle size and surface area are limiting

step for dissolution kinetic of the API release rate out of this formulation.

Our goal is to design and validate a robust dissolution method that would be able to provide

reliable and accurate data of the API-release profile from the pharmaceutical formulation in in

vitro conditions. Such method would be valuable in a generic drug development or the

enhancement of the formulation in a personalized medicine.

RP-HPLC METHOD - DEVELOPMENT & VALIDATION

An integral part of the dissolution method is the analytical method suitable for determining API

and prodrug concentrations. Since a large number of substances are present in the dissolution

mixture (prodrug, API, suspension stabilizers, buffer, etc.), the individual components must be

separated from each other prior to the analysis. For this purpose the High Performance Liquid

Chromatography (HPLC) technique was chosen. As there was no HPLC method for the parallel

determination of API and prodrug concentration, it was necessary to develope a new one. The

design and validation of this method are described in the following paragraphs.

Instrumentation & Chromatographic Conditions

The analysis was carried out by using Agilent 1100 HPLC system equipped with quartenary pump,

autosampler, degasser and UV-Visible detector. The system was monitored by using OpenLab

software. HPLC separation was carried out on a Symmetry C18 column (75 mm × 4.6 mm; 3.5


µm) maintained at 37°C with a mobile phase consisting of 10 mM ammonium phosphate buffer

(pH 2.5) and acetonitrile with a gradient flow described in Tab. 1. A volume of 10 µL was injected

and all the analytes were monitored at 279 nm. The run time for the analysis was 10 min.

Tab.1 – gradient elution of the HPLC method

Time [min] Buffer [vol. %] ACN [vol. %] Flow Rate [mL/min]

Initial 90 10 1.000

4.00 50 50 1.000

4.50 20 80 1.000

7.00 20 80 1.000

7.50 50 50 1.000

10.00 90 10 1.000

Standard Solutions & HPLC Calibration

The mixed standard stock solution with the concentration of 126 mg/mL for API and 147 mg/mL

for prodrug was prepared. Both substances were dissolved in the solvent consisting of Tween20,

THF and phosphate buffer (same as for mobile phase). By dilution of the stock solution a

concentration row was created and subsequently used for calibration. The chromatogram

belonging to the mixed standard solution is shown in Figure 1. It can be seen that both substances

are well separated from each other and the peaks are Gaussian in shape. The measurement results

of the signal response dependency on the concentration of substances are illustrated in Figure 2.

Concentrations of both substances are within a linear range and the correlation coefficients meet

the limits for method validation - R ≥ 0.999 [3].

Methodology of Sample Preparation

The dissolution mixture was consisting of suspension of prodrug, stabilizers, buffer, antibiotics,

enzyme and the dissolved API as a product of reaction. One liter of such mixture was put into the

glass container that was placed into a shaker-incubator (37°C, 240 rpm). To determine the

0

50

100

150

200

250

0 2 4 6 8 10

Sign

al s

tren

gth

[m

AU

]

Time [min]

0

1000

2000

3000

4000

5000

0 50 100 150 200

Are

a

Concentration [mg/l]

API R = 0.999991

Prodrug R = 0.99995

Figure 1: Chromatogram of mixed

standard solution

Prodrug

API

Figure 2: Calibration model for

quantitative analysis


concentration of released API, samples of 5 mL were taken and boiled for 10 min in order to stop

the reaction. Then the samples were centrifuged (13,400 rpm, 20 min) and the solution above a

solid phase was transferred into HPLC vial. This solution was then measured by the new HPLC

method.

Recovery Factor

This factor is an important part of the the analytical method validation. For the drug product, its

evaluation is performed by the addition of known amounts of drug by volume to the formulation

working in the linear range of detection. This test evaluates the specificity of the method in the

presence of the excipients under the chromatographic conditions [3].

In order to evaluate recovery, samples of the dissolution mixture were taken after 99 hours from

mixing the suspension with the enzyme solution (beginning of the enzymatic reaction). The

samples were boiled to stop the reaction and the method of standard additions was made for both,

crude and milled suspension. The results are shown in Figures 3 and 4 where on the x axis there is

a known concentration addition of API and on y axis there is a measured concentration of API by

HPLC. The recovery factor is then the slope of the regression multiplied by 100%. To validate the

method, it is recommended for the RF value to be between 90 and 110% [3].

Figure 3: Recovery factor of crude suspension

Figure 4: Recovery factor of milled suspension

API-RELEASE PROFILE

Using the proposed HPLC method, kinetics of the enzymatic reaction was measured in a solution

of a prodrug with varying enzyme concentration. The results are shown in Figures 5 and 6. It can

be seen that the concentration range of the prodrug (substrate) in solution falls to the left of the

Michaelis-Menten plot. As a result, there is no acceleration of the reaction after a certain enzyme

concentration has been exceeded. Due to the very low solubility of the prodrug, it was not possible

to measure the reaction rate for a given medium for higher substrate concentrations. This is also

the reason why the M-M kinetic constants could not be determined.

y = 0.9884x + 7.1215 R² = 0.9998

0

20

40

60

0 20 40 60

Mea

sure

d c

(AP

I) [

mg/

l]

Known Addition of API [mg/l]

RF = 98.8%

y = 1.0010x + 5.9839 R² = 1.0000

0

20

40

60

0 20 40 60

Mea

sure

d c

(AP

I) [

mg/

l]

Known Addition of API [mg/l]

RF = 100.1%


Figure 5: Dependency of enzymatic reaction

rate on enzyme concentration

Figure 6: Fitting the data from Figure 3 to the

Michaelis-Menten model

100 ENZ = used concentration of enzyme was 100 mg/l (analogically for other numbers)

r(E100) = rate of enzymatic reaction for enzyme concentration 100 mg/l (analogically for other numbers)

The following chart in Figure 7 shows the results from the measurement of the enzymatic

dissolution from the prodrug suspensions. It can be seen that with a small concentration of

enzyme, the API growth rate is the same for the crude and milled suspensions. When increasing

the enzyme concentration, the release rate of API from the suspensions differs. The magnitude of

the API concentration increase is greater for the milled suspension since the limiting step in this

case is the dissolution rate that is proportional to the surface area of the solid particles.

Figure 7: Reactive enzymatic dissolution rate

CONCLUSION

The dissolution test for measuring API-release profile has been proposed. Within this test, the

HPLC method for parallel determination of the API with the prodrug and the methodology of

samples preparation were designed. The HPLC method meet the limits for validation, however

due to the insufficient repeatability of overall process further investigation of the methodology is

needed.

REFERENCES [1] Jeremy, W. C.; Diane, B. J., Long Acting Injections and Implants. Springer: Boston, 2012. [2] Andersson, S. B. E.; Alvebratt, C.; Bergström, C. A. S., Controlled Suspensions Enable Rapid Determinations of Intrinsic

Dissolution Rate and Apparent Solubility of Poorly Water-Soluble Compounds. Pharmaceutical Research 2017, 34 (9), 1805-1816.

0

5

10

15

20

25

0 10 20 30 40

Co

nce

ntr

atio

n o

f A

PI [

mg/

l]

Time [days]

100 Enz

50 Enz

10 Enz

0 Enz

0

0.01

0.02

0.03

0.04

0.05

0.06

0 10 20 30 40 50 60 70 80 90

dc(

AP

I)/d

t [

mg/

(l·h

)]

Concentration of IM (substrate) [mg/l]

r (E10)

r (E50)

r (E100)

0.0

1.0

2.0

3.0

4.0

0 2 4 6 8 10

Co

nce

ntr

atio

n (

AP

I) [

mg/

l]

Time [days]

median particle diameter 13 µm; c(enzyme) = 90 mg/l






[3] Reviewer Guidance, Validation of Chromatographic Methods - FDA; Center for Drug Evaluation and Research, 1994.


Design of Taylor reactor for continuous preparation of silica microparticles

Bc. David Zůza

školitel: Ing. Ondřej Kašpar, Ph.D.

školitel – specialista: Ing. Marek Šoltys

Keywords: silica particles, Taylor-Couette Flow, mesoporous, 3D printing, CFD

1. Introduction

1.1. Silica particles

Silica particles, both nano- and micro-sized, have been in the scope of many research groups for their

tuneable size, morphology, and high thermal and mechanical stability. Silica particles can be fabricated in various

shapes, e.g., spherical1, rod-like or cubic and their size and size distribution can be controlled by a chosen method

of fabrication2. Moreover, the possibility of their additional surface modification makes them a promising

candidate for a wide range of pharmaceutical and therapeutic applications. Mesoporous silica materials, discovered

in 1992 by the Mobile Oil Corporation, attracted a lot of attention due to exceptionally high values of specific

surface and well-ordered pores (2 to 50 nm) with the ability to accommodate even larger cargo molecules3.

One particular application is the use of silica particles in oral drug delivery formulations. Particles larger

than 100 nm can be safely used in oral delivery applications since particles of this size are no longer passing

through the intestinal wall and are safely excreted from the gastrointestinal tract. There is a number of reported

techniques and protocols how to produce silica. However, emulsion methods based on hydrolysis of Tetraethyl

orthosilicate (TEOS)1 are the most commonly implemented in the preparation of mesoporous silica particles. Using

emulsion technique, the mesoporous structure can be controlled by the presence of surfactants.

1.2. Taylor – Couette reactor/flow

Taylor-Couette reactor (TCR) is a batch reactor made from two concentric rotating cylinders separated

by fluid (Figure 1A). Cylinders can rotate in opposite or in the same direction, which results in eighteen mixing

regimes described in the Figure 1B4 it is called Taylor-Couette flow. In real application modification with the

steady external cylinder is mostly used. This modification provides only five discovered flow regimes according

to the literature4 (Couette flow, Taylor vortex flow, Wavy vortex flow, Modulated waves flow, and turbulent Taylor

vortices flow). TCRs are in general employed for homogenous mixing of highly viscous mixtures such as

substrates for enzymatic catalyzes, reactions, and also for preparation of submicron particles (lower viscosity),

however, their feasibility for microparticle synthesis still needs to be investigated. In this work continuous TCR

(cTCR) is designed and developed and the possibility of creating narrower particle size distribution by applying a

narrower shear rate distribution provided by TCR especially by Couette flow regime is investigated.

1.3. 3D printing

3D printing is a method of fabrication of 3D structures from various types of materials (e.g., polymers

like PLA (Polylactic acid), ABS(Acrylonitrile butadiene styrene)) based on a virtual 3D model. The most common

3D printers employ the Fused Deposition Modelling (FDM) technology, which extrudes material heated to the

melting point through an orifice of printing head nozzle and draws planar layer from the extruded material. The

desired 3D object is printed layer by layer in the z-axis. Another commonly used 3D printing method is

Stereolithographic printing (SLA). Solidifying of photopolymer layer by layer by UV light is used for object

fabrication. The possibility to print almost any shape and to create sophisticated inner structures are some of the


most significant benefits of 3D printing technology. Nowadays, it can be used anywhere from printing houses,

living tissues containing cells for implants, to the fabrication of micro-reactors or micro-reaction systems as

microfluidic devices with a resolution smaller than 0.1 mm5.

A)

B)

Figure 1: A) Scheme of Taylor-Couette flow in TCR. B) The figure of turbulent regimes produced by two rotating cylinders

R0 (x-axis) and R1 (y-axis), because TCR with steady outer cylinder is used R0 is equal to 0, which reduces expected flow

regimes on five4.

2. Experimental

2.1. Materials

Tetraethyl orthosilicate (TEOS, >99%) and octylamine (OA, >99%) were purchased from Sigma-Aldrich,

ethanol (absolute, 99.9%) were purchased from Penta Chemicals. All chemicals were used as received without

further purification. Demineralized water (Aqual 25, conductivity of 0.07 μS/cm) was used for all reactions and

treatment processes.

2.2. Batch Preparation of silica particles

A mixture of 10 ml of OA and 10 ml of TEOS was prepared in a 200 ml PTFE beaker and stirred at 800

rpm (25x5 mm stirring rod) for 3 minutes by room temperature. Then, 133 ml of ethanol-water (1:3, v:v) mixture

was added and stirred for 3 minutes at the same conditions. The collected particles are then centrifuged for 5

minutes at 3740 rcf, redispersed in the ethanol-acetone mixture (1:1, v:v) and centrifuged. This washing process

was repeated two times. The silica particles were dried in a crystallization bowl and calcined. Calcination set up

was 6 hours ramp from room temperature to 600 °C and then kept for 6 hours at 600 °C.

2.3. TCR preparation of silica micro particles

2.3.1. TCR1

TCR1 is a continuous TCR, where inner cylinder has a diameter 3 cm, outer cylinder has diameter 5 cm and

height of the reactor is 2 cm. Inflow is separated for a premixed mixture of ethanol and water and for a premixed

mixture of OA and TEOS. Both inflows are set up as described in Figure 2.

Preparation of the particles using TCR1 is done by linear pumping of both mixtures in the ratio given by batch

preparation, and final flow rates are calculated in respect to the residence time of 3 minutes. The flow rate for

OA/TEOS mixture is 1 ml/min, and flow rate for EtOH/water mixture is 5 ml/min. Revolutions used for inner

cylinder were 400, 600, 800 and 1000 rpm.

2.3.2. TCR2

TCR2 is a continuous TCR, where inner cylinder has diameter 2 cm, outer cylinder has diameter 3 cm and

height of the reactor is 5 cm, in contrast to TCR1 where dimensions were 3, 5 and 2 cm, respectively ratio

gap/height to be 1/10. In comparison to TCR1, the main change is in the gap/height ratio (1/21/10) which

governs flow and shear distribution. Inflows are set up as described in Figure 2.

Studied region

R0 =0


Preparation of the particles in TCR2 is realized by pumping of mixture OA/TEOS and mixture EtOH/water

in the ratio given by batch preparation. The final flowrates are adjusted in such a way that mean residence time

inside of the reactor is 3 minutes. The flow rate for OA/TEOS mixture is 1.3 ml/min, and flow rate for EtOH/water

mixture is 6.5 ml/min. The inner cylinder was rotated at 400, 600, 800 and 1000 rpm.

Figure 2:3D visualization of both reactors and cross-section 2D visualization.

2.4. Modeling of Taylor-Couette (TC) flow

COMSOL Multiphysics (CFD program) was used for the study of fluid flow inside of TCR. Simplified reactor

geometries without reactant inlets were used to reduce computational time and complexity of the solved problem.

Properties of liquid such as density and viscosity were approximated by corresponding EtOH/water mixture.

Deformation of free surface was approximated by the symmetric boundary condition. The no-slip boundary

condition (fluid flow velocity will be near the surface zero relative to the boundary) was imposed on the outer wall

and on the floor of the reactor. Gravity as the external force was included due to macroscopic size of the solved

system. Turbulent model k-ε (kinetic and dissipation energy relation) was preferably used, because of its high

stability and robustness. Velocity and shear rate distribution was studied for the same conditions used during

experiments.

3. Results and discussion

3.1. Comparison of prepared particles

Particles from batch preparation (Figure 3A) were spherical, hollow-core and multi-structured from submicron

particles (Figure 3B). Their particle size distribution (PSD) was ranging from 10 to 70 µm (Figure 3D). In an effort

to pick up the best samples from cTCR1, their PSDs and morphologies were compared. The most promising results

were obtained with an inner cylinder rotating at 1000 rpm. Particles prepared from cTCR1 (1000 rpm) were

spherical, multi-structured and hollow-core (Figure 3C). Their PSD is in the range of 5 to 70 µm (Figure 3D).

Consequently, the best rotational speed for cTCR2 in terms of morphology and particle size was 400 rpm.

Particles were also spherical, multi-structured and hollow-core (Figure 3E), but their surface was covered with

unstructured submicron silica particles more than was usually observed in batch or cTCR1 preparation. PSD was

in the range from 2 to 60 µm (Figure 3D). In respect to particle quality (i.e., size, morphology, PSD) particles

prepared by cTCR1 at 1000 rpm were superior to those produced by cTCR2.

A)

B)

C)


D)

E)

Figure 3: A) SEM photo of particles prepared in batch. B) TEM photo of submicron compartments of batch particles. C) SEM

photo of particles prepared in cTCR1. D) Comparison of PSDs prepared particles. E) SEM photo of particles prepared in

cTCR2.

3.2. CFD simulation

Models were calculated as described in section 2.4. TC vortices were obtained in both cases. In the case of

cTCR1 model, 1 or 2 vortices are observed in dependence on applied rotational speed and time (Figure 4A). In

cTCR2 4 or more vortices were obtained for various rotational seep and time (Figure 4B). For comparison, time-

averaged distributions of shear rate were created. From obtained results for cTCR1 (1000 rpm) and cTCR2 (400

rpm) (Figure 4C) can be concluded influence of geometry on flow character. The width of shear rate distribution

is similar for both cases, but for cTCR1 we can see the majority of data points in the region from 5 to 50 s-1. This

fact can explain particles covered with submicron silica and debris obtained from cTCR2. It can be expected that

some particles are not able to withstand high shear stress in some regions of the reactor which results in their

collision, attrition, and fragmentation.

From previously described results can be concluded that the lower rpm can be the solution to the problem

with the non-smooth surface of particles produced by cTCR2, which are undesirable, because of our main goal

particle uniformity.

A)

B)

C)

Figure 4: A) Calculated velocity stream lines profile in cTCR1(1000 rpm) B) Calculated velocity stream lines profile in

cTCR2 (400 rpm). C) Shear rate distribution comparison.

4. References

1. Zůza, D.; Šoltys, M.; Mužík, J.; Lizoňová, D.; Lhotka, M.; Ulbrich, P.; Kašpar, O.; Štěpánek, F., Silica particles with three levels of porosity for efficient melt amorphization of drugs. Microporous and Mesoporous Materials 2019, 274, 61-69. 2. Zhang, K.; Gan, L. M.; Chew, C. H.; Gan, L. H., Silica from hydrolysis and condensation of sodium metasilicate in bicontinuous microemulsions. Materials Chemistry and Physics 1997, 47 (2), 164-170. 3. Bharti, C.; Nagaich, U.; Pal, A. K.; Gulati, N., Mesoporous silica nanoparticles in the target drug delivery system: A review. International journal of pharmaceutical investigation 2015, 5 (3), 124. 4. Andereck, C. D.; Liu, S. S.; Swinney, H. L., Flow regimes in a circular Couette system with independently rotating cylinders. Journal of Fluid Mechanics 2006, 164, 155-183. 5. Ma, H.; Feng, C.; Chang, J.; Wu, C., 3D-printed bioceramic scaffolds: From bone tissue engineering to tumor therapy. Acta Biomaterialia 2018, 79, 37-59.

0

5

10

15

20

1 10 100 1000

Vo

lum

e f

ract

ion

(%

)

Particle diameter (µm)

Batch 800 rpm

cTCR1 1000 rpm

cTCR2 400 rpm

0%

2%

4%

6%

8%

0.5 5 50 500 5000

Nu

mb

er

frac

tio

n

(%)

Shear rate (s-1)

cTCR2 400 rpmcTCR1 1000 rpm


Linking micro-scale and meso-scale models for catalytic filter

Martin Šourek

supervisor: doc. Ing. Petr Kočí, Ph.D.

supervisor-specialist: Ing. Martin Isoz, Ph.D.

Keywords: CFD, Catalytic filter, Filtration, OpenFOAM

Introduction

The development of automotive industry and increasing number of vehicles is causing

environmental and health issues [1]. Soot particles present in automotive exhaust gas are serious

problem leading to increasingly stringent automotive emission regulations [2]. This enforces

the use of particulate filters to effectively control the particulate emissions from both diesel and

gasoline engines [3].

Presence of a particulate filter further increases the size and complexity of the

automotive exhaust gas-after treatment. To make the exhaust system more compact, the

particulate filter can be combined with catalytic convertors. By depositing the catalytic material

directly into the filter a new device, a catalytic filter, is obtained [4].

In this work, I describe a reliable multi-scale CFD model for catalytic filters and suggest

an improvement to this model based on a better inclusion of macro-scale data from simulation

of the whole channel into micro-scale computations. The developed model relies on

OpenFOAM, an open-source library for computational continuum mechanic [5]. A scanned

structure of a catalytic filter is transformed into a computational mesh by image processing and

snappyHexMesh, the OpenFOAM utility for meshing of complex geometries. The fluid flow

through the porous media is computed using the solver porousSimpleFoam from the

OpenFOAM library. The soot filtration process is solved by a custom Lagrangian particle-

tracking solver implemented in OpenFOAM.

At the moment, there is only a one-way coupling between the micro- and macro-scale

models, i.e. the data computed on micro-scale are used for macro-scale simulations but not the

other way around. As a result, the flow field in the micro-scale model is unrealistic and the

model tends to under-estimate the catalytic filter filtration efficiency. To improve the prediction

of micro-scale flow field, I developed a new boundary condition for mapping the flow data

from macro-scale to micro-scale models. The new boundary condition enables a correct two-

way coupling between the different scales of the overall multi-scale model.

Model description

Geometry

One sample of a real catalytic filter is examined in this study. The sample is based on a

cordierite filter coated with Al2O3 in a way that the coating is present only in the porous wall.

First, a detailed description of the porous filter with distributed catalytic material is

needed. For this purpose, x-ray tomography is utilized to scan partitions of the catalytic filter

with resolution of 2016 x 2016 pixels and size of one voxel 1.1132 μm. Image processing is

used to transform scans into STL files and to obtain a digital representation of pores

morphology. Then, computational mesh based on STL file is created with snappyHexMesh.

Note that pores of the filter are meshed as a free space and the coating as a porous zone. The


meshing process itself is fully automatic. The output of snappyHexMesh is an unstructured

mesh conforming to the generated porous geometry.

The simulated system is represented in model equations as follows: 𝛺 ⊂ 𝑅3 is a simply

connected open domain including both the free pores (𝛺𝑃) and the coated zones (𝛺𝑐), such that

𝛺 = 𝛺𝑃 ∪ 𝛺𝑐 and 𝛺𝑃 ∩ 𝛺𝑐 = ∅, and Γ is the boundary of the domain 𝛺.

CFD Setup

A steady-state isothermal flow of an incompressible fluid is considered. The fluid flow

was simulated in both open pores (𝛺𝑃) and coated zones (𝛺𝑐) simultaneously. The Navier-

Stokes equations governing the flow are,

∇ ∙ (𝑢 ⊗ 𝑢) − ∇ ∙ (𝜈∇𝑢) = −∇𝑝 + 𝑠

∇ ∙ 𝑢 = 0,

where 𝑢 stands for the velocity field, 𝑝 for the kinematic pressure and the coefficient 𝜈 denotes

the fluid kinematic viscosity. All the external body forces exerted on the fluid are neglected

The additional source term 𝑠 in equation (1) is obtained from the Darcy permeability

model and it is defined as,

𝑠 = {0 ∈ ΩP

−𝜈

𝜅𝑐𝒖 ∈ Ωc ,

where 𝜅𝑐 is the local Darcy permeability through the coated catalytic medium. The Darcy

permeability of the coated material is estimated from the Carman-Kozeny equation as

2.76 ∙ 10−16 m2 which equals to 0.00276 Da.

The system (1) needs to be supplemented with suitable boundary conditions. In the

original model, the boundary of the solution domain was divided into three sections, inlet, outlet

and walls. A uniform velocity in the direction orthogonal to the channel wall and a zero-gradient

boundary condition for the pressure were prescribed at the inlet. Constant pressure and inlet-

outlet boundary condition for velocity were specified at the outlet. A standard no-slip boundary

condition was prescribed at walls.

Micro-scale and macro-scale link

The boundary conditions in the original model do not take into account the actual flow

field in channels of the catalytic filter. The flow close to the channel wall is seldom orthogonal

to it. In fact, most of the gas commonly flows along the channel and not across the wall. Thus

the estimated micro-scale velocity field is unrealistic. Furthermore, the estimated device

filtration efficiency is under-estimated.

To improve the boundary conditions, the following is proposed:

1. Compute the flow field in micro-scale model using the unrealistic inlet velocity. Use

the result to estimate the permeability of the catalytic filter wall.

2. Compute the flow field in the complete monolithic catalytic filter using the macro-scale

model. Approximate the channel walls as porous zones with permeability obtained from

1. To save the computational resources, leverage the device periodicity and simulate

only a single representative unit of it.

(1)

(2)


3. Map the flow field obtained from 2 to the boundaries of the micro-scale model. Compute

realistic velocity field in the channel wall.

4. Use the velocity field from 3 as input for Lagrangian simulations of soot particles

filtering.

Note that the fact that the wall permeability does not depend on the overall flow

direction with respect to the wall enables us to compute each of the above steps only once.

The main contribution of this work is development of an approach to perform the step

3 of the above outlined process. First, it is necessary to collect relevant velocity data from the

macro-scale model. This is done in the Paraview software as a postprocessing to the macro-

scale simulations. The (planar) boundaries of the micro-scale model are projected into the

macro-scale geometry. For each plane corresponding to the micro-scale boundary, we save the

available velocity data as 𝑢ℎ = 𝑢ℎ(𝜁1, 𝜁2), where (𝑂, 𝜁1, 𝜁2) is the local coordinate system

defining the processed plane. The obtained data are discrete and noisy due to numerical errors.

The 2D least-square method is used to fit a polynomial to the data and to obtain a smooth and

continuous representation of the velocity field on the processed plane. The obtained

polynomials are scaled to provide a consistent velocity field and used to define the boundary

conditions for the micro-scale model.

The specification of the new boundary conditions and the mapped velocity field is

depicted in Figure 1. To show all the newly defined boundaries, two different views of the

same structure are given. Note that in order to have a well-defined problem, it is necessary to

prescribe the outlet-inlet and zero-gradient boundaries for velocity and pressure, respectively

at one of the patches (outletBot in our case).

Figure 1. The specification of new boundary conditions and the mapped velocity field


Results

In order to verify suggested approach, the soot filtration for both boundary conditions

is computed and the filtration efficiency is evaluated. Three different positions in macro-scale

simulation are used for velocity mapping because the velocity profile changes along the

channel. These positions are namely: inlet, centre and outlet of the channel. The filtration

itself is computed using Langrangian approach. The filtration efficiency is evaluated based on

the information about the number of particles trapped in the porous wall and passed through

the wall. Results are depicted in Figure 2.

Figure 2. Comparison of filtration efficiency for both boundary conditions at different channel

positions.

Conclusion

The original micro-scale flow model as published in [6] provides an unrealistic velocity field

in the microstructure of the catalytic filter wall. Consequently, the filtration model, which

uses a computed velocity field as one of the inputs, under-estimates the filtration efficiency of

the catalytic filter. I have proposed a method for improvement of the flow model based on

linking the micro-scale and macro-scale simulations of the catalytic filter. Micro-scale model

with the new boundary conditions gives more realistic prediction of velocity profile and better

results for filtration efficiency compared to the original one.

References

1. G. Oberdörster, M.J.U., Ultrafine particles in the urban air: to the respiratory track

and beyond. Environmental Health Perspectives, 2002. 110: p. A440-A441.

2. Johnson, T., Vehicular Emissions in Review. SAE Int. J. Engines, 2016. 9(2): p. 1258-

1275.

3. Lambert, C., Chanko, T., Dobson, D. et al., Gasoline particle filter development.

Emission Control Science and Technology, 2017. 3(1): p. 105-111.

4. Y. Zheng, Y.L., M.P. Harold, D. Luss, LNT-SCR dual-layer catalysts optimized for

lean NOx reduction by H2 and CO. Applied Catalysis B: Environmental, 2014. 148-

149: p. 311-321.

5. OpenFOAM. https://openfoam.com/. 10.11.2018].

6. Kočí, P., et al., 3D reconstruction and pore-scale modeling of coated catalytic filters

for automotive exhaust gas aftertreatment. Catalysis Today, 2019. 320: p. 165-174.

0

10

20

30

40

50

60

Inlet Center Outlet

Filtration efficiency

[%]

Original BC New BC

https://openfoam.com/

Protocells: Promising nanocarriers for on-demand drug ...uchi.vscht.cz/uploads/svk/E1.pdf ·...

Documents

Transcript of Protocells: Promising nanocarriers for on-demand drug ...uchi.vscht.cz/uploads/svk/E1.pdf ·...