Purification and repossession of methanol in a polymer synthesis process by fractional distilling

(1)

LUT School of engineering science Master’s Programme in Biorefineries

Examiners: Professor, D. Sc (Tech) Tuomas Koiranen Assistant professor, D. Sc (Tech) Kristian Melin

Delivery capability manager, M. Sc (Tech) Juulia Loisa Vesa Leppänen

PURIFICATION AND REPOSSESSION OF METHANOL FROM A POLYMER SYNTHESIS PROCESS BY FRACTIONAL DISTILLING

(2)

ABSTRACT

LAPPEENRANTA-LAHTI UNIVERSITY OF TECHNOLOGY LUT LUT School of engineering science

Master's Programme in Biorefineries Vesa Leppänen

Purification and repossession of Methanol in a polymer synthesis process by fractional distilling

Master's thesis 2021

79 pages, 32 figures, 12 tables, and 11 appendices Examiners: Professor, Dr. Sc (Tech.) Tuomas Koiranen

Assistant professor, Dr. Sc (Tech) Kristian Melin Delivery capability manager, Ms. Sc (Tech) Juulia Loisa

Keywords: solvent recovery, fractional distilling, solvent recycling, methanol purification ___________________________________________________________________________

The purpose of this Master's Thesis was to create a new process protocol for solvents repossession from a polymer synthesis process and to recycle the gained material back into the process in the future phases of the synthesis. This thesis aimed to decrease solvent consumption by refining the material by fractional batch distillation to reduce the quantity of chemical waste and overall expenses, including acquiring and disposing of the chemicals. Basics of fractional distillation were reviewed, including distillation equipment in the literature part. In addition, commercial equipment was reviewed.

A new process protocol is based on the study and the results presented in this thesis. Overall, the study is based on a theoretical review of the chemicals and purification method—the experimental part and the presentation of the purification process results. The success is evaluated upon the reached purity levels and the cost reductions of laboratory management and waste. This thesis was limited to handling only fractional distilling because of the available equipment.

The gas chromatography analysis indicates that reached purity of the methanol sample in the final stage is acceptable. Based on these initial test results, we can engineer a new protocol for solvent

(3)

repossession in the polymer process. The new process protocol will decrease the consumption and waste inflicted by the current polymer process. In addition, the chemical acquiring, and waste management costs will decrease in the long run, which will cut costs affiliated with the polymer process.

(4)

TIIVISTELMÄ

LAPPEENRANNAN-LAHDEN TEKNILLINEN YLIOPISTO LUT LUT School of engineering science

Kemiantekniikka Vesa Leppänen

Metanolin puhdistus ja takaisinotto polymeerisynteesiprosessista fraktionaalisella tislauksella Diplomityö

2021

79 sivua, 32 kuvaa, 12 taulukkoa ja 11 liitettä

Tarkastajat: Professori, D. Sc (Tech.) Tuomas Koiranen Apulaisprofessori, D. Sc (Tech) Kristian Melin

Delivery capability manager, M. Sc (Tech) Juulia Loisa

Hakusanat: liuottimen talteenotto, fraktionaalinen tislaus, liuottimen kierrätys, metanolin puhdistus

___________________________________________________________________________

Työn tarkoituksena oli suunnitella uusi prosessi liuottimien talteenotolle polymeerisynteesiprosessista. Talteenotettu liuotin on tarkoitus jatkokäyttää uudelleen synteesiprosessin eri vaiheissa. Tämän työn pohjalta oli tarkoitus pienentää kemikaalien kulutusta jatkojalostamalla ne käyttöön uudelleen. Puhdistus tehdään fraktionaalisella tislaimella. Liuottimien puhdistus ja jatkojalostus vähentää kemikaalien kokonaiskulutusta. Kulutuksen pienentämäisellä saavutetaan taloudellisia sekä ympäristöllisiä parannuksia. Jatkokäyttö vähentää uusien kemikaalien hankintakuluja sekä käytettyjen kemikaalien jätekustannuksia. Lisäksi, ympäristökuormitus vähenee sekä suoraan että välillisesti.

Työ perustuu kirjalliseen sekä kokeelliseen osaan. Kirjallisessa osassa käydään läpi fraktionaalisen tislaimen laitteisto sekä sen toiminta. Perustana kirjalliselle osalle on tätä tarkoitusta varten hankittu laitteisto. Kirjallisuusosassa on esitelty teoreettinen tarkastelu tislauslaitteiston suunnittelulle ja tarkasteltu tässä työssä käytettävät kemikaalit. Kokeellinen osa koostuu analyysilaitteiden esittelystä sekä työssä suoritettujen kokeellisten testien parametrien esittelystä sekä parametrien varioinnin tuloksista. Tämä työ on rajoitettu koskemaan ainoastaan fraktionaalista tislausta sekä metanolin puhdistusta.

(5)

NOMENCLATURE

Greek symbols

ij interaction energy between i and j

x activity coefficient of a substance x, dimensionless

Roman symbols

B bottom product flow, mol/s

C number of chemical components, dimensionless F degrees of freedom, dimensionless

L total flow rate of liquid, mol/t n number of moles, mol

P number of phases present, dimensionless p pressure, Pa

pi partial vapor pressure, Pa

Pi* partial vapor pressure, pure substance, Pa t time, s

T temperature, K V volume, m³

(6)

V total flow rate of vapor, mol/t xi mole fraction in the mixture, mol xD mole fraction in the distillate, mol

xB mole fraction in the bottom product, mol yi molar fraction in vapor, mol

(7)

1. Introduction

1.1.

Background and objectives of the thesis

Initial prestudy showed that the production quantities in the production line in question had increased steadily over the recent years. The production increase has affected polymer consumption, and alongside solvent consumption. Increasing material consumption in production requires an update and upscaling of the polymer synthesis process. The synthesis process uses different solvents, and the more extensive synthesis process requires a larger quantity of these solvents. The more significant consumption has created a need to recycle the solvents to cut back costs, waste and make the process more sustainable. This thesis studies the purification of these solvents, N-methyl-2-pyrrolidone, and Methanol, focusing on the latter. The method for purification used in this thesis is fractional distilling.

This thesis aims to purify solvents via fractional batch distillation, a popular solvent purification method. These devices are readily available on the market, and the one used in this research is commercially available. A comparison of the lab-scale distillation systems, which were considered for the purchase, is handled in the literature review. The solvents, which are purified, will be put back into use, either in the synthesis or other necessary operations in the laboratory. The purity of the distilled solvents will be verified by different analysis methods in the experimental section. In addition, this research aims to reach a desired level of purity in repossessed solvents. The final goal is to have a repossession process in regular production after this process development.

(10)

1.2.

Research methods and outline of the thesis

The study will take place in two segments: a literature review of the basics of distillation and an experimental part. In the first part, there is a review of the purification method and the materials. The latter section will focus on empirical testing. We will acquire the waste solvents organically from the synthesis process; instead of disposing of the solvents, we will receive these as test material. The collected waste material is then purified with the fractional distillation device and analyzed by gas chromatography and Karl-Fischer titration.

There are different ways to purify this waste material, but the thesis's outline considers only fractional batch distillation. This limitation is in place because of available instrumentation. In addition, the selected solvents for the purification process are the most consumed in this synthesis process, contributing the most to the waste quantity. Therefore, the thesis is limited to considering only these solvents.

2. Basics of fractional distillation

Distillation can be defined as a separation of substances drop by drop. The separation principles have been used for thousands of years, yet the modern distillation process is only roughly 150 years old. The term distillation describes the separation of two similar substances with different material qualities by evaporation and condensation. (Krell, 1982)

(11)

2.1.

Phase diagram

Each substance has an individual phase diagram. The graph tells us about substances' behavior and phase changes. An example substance undergoes each phase shift; solid-liquid, liquid-gas.

If the lines were to cross in a different series, e.g., the gas line would cross solid before liquid, we could conclude that the substance sublimates. The phase transition temperature is also one thing to consider. The phase shift does not happen straight away but rather slowly while there are usually two different phases. For example, when water begins to boil at 100°C (atm), both the liquid and the gas phase are present. (Campbell, 2012)

The phase diagram aims to illustrate a substance's most stable thermodynamic phases at different temperatures and pressures. There are a few terms, which are crucial to understanding the phase diagrams:

➢ Triple point – The point where all three phases, solid, liquid, and gas, can be present at the same time

➢ Critical point – The point where gaseous and liquid phases cannot be identified

➢ Fusion curve (melting or freezing curve) – Transition between liquid and solid-state

➢ Vaporization curve – Transition between liquid and gaseous state

➢ Sublimation curve – Transition between solid and gaseous phase

These graphs are usually presented as pressure as a function of temperature, see Figure 1.

(Zhang, Zhang, et al., 2018)

(12)

Figure 1. An example of a phase diagram. There is a lot of information on the substance in one Figure. (Zhang, Zhang, et al., 2018)

The fusion curve (green) is for most substances going from left to right as water makes an exception to this phase diagram (dotted green). The green line marks the freezing curve, and the blue line marks the boiling curve to simplify the Figure. We have a distinct and essential concept at the phase boundaries considering the distillation process – vapor pressure. The vapor pressure results from the molecules leaving their liquid phase and causing the increase in pressure. Vapor pressure depends heavily on temperature, as the molecules require more energy to escape the initial state. (Atkins, de Paula. 2009)

In the distillation process, two or more equally soluble liquids are separated by evaporation and condensation. Distillation does not define the number of materials involved in the process; hence, if we evaporate and condense water, it is also distilling, although no actual

(13)

separation occurs. Simple distillation is the generally understood process when considering distillation. However, simple distillation cannot achieve desired purities in the actual separation process and is relatively slow and expensive to use outside the laboratory scale.

(Krell, 1982)

In comparison, batch distillation and continuous distillation with multiple stages are more convenient and efficient. We utilize a batch distillation with multiple stages in this thesis;

hence the continuous process is neglected. There are minor differences between the continuous and batch processes, and it is mainly capacity dependent, which one is more feasible. Figure 2. shows a schematic view of a fractional batch distillation system used in this thesis. (Divekar, 2011)

Figure 2. Schematic drawing of batch distillation device with multiple stages. (Gorak, Sorensen, 2014)

(14)

The overall assembly is relatively simple, as seen in figure 2. The device consists of:

➢ Reboiler – the material is heated to its boiling point in the reboiler

➢ Fractionating column – filled with thermodynamical stages

➢ Condenser – condensing the vapor to liquid

➢ Reflux – returns desired fraction of the liquid back to the distillation column

➢ Material collection system – distilled material is led in separate flasks

There are several factors to consider in the distilling process, e.g., the properties of the substances, boiling points, column, and equipment structure. Simple distillation is feasible if the materials have significant differences in boiling points, over 50°C, respectively. If the boiling points are closer, fractional distillation is required. The fractional distilling can separate materials close to the same boiling point (0.05°C – 0.5°C). The effect of the fractional column is dependent on chosen materials and column types. In this thesis, the distillate composition is unknown, but we know that the majority is Methanol. (Krell, 1982) We have acquired this specific fractional distillation equipment because we don't know the feed composition and the future development possibilities.

2.2.

Distillation system

Here, we consider the different parts of the distillation system. The components are limited to laboratory equipment. This thesis does not consider industrial scale, as it is not in the scope nor of interest for us.

(15)

2.2.1. Reboiler

In the distillation system, the material must be heated to vaporize material and achieve separation, depending on material-specific boiling points. There are a few different types of reboilers, and it is up to the final process, which is the best available option. The few common reboilers to mention are jacketed kettle reboiler, thermo-siphon reboiler, and internal boiler.

The latter ones are more common in large-scale industries, such as oil stripping, and are in no interest for us. In our system, the reboiler is the jacketed heater-type. (Tham, 2016) The function is relatively simple since it is a resistive heating mantle for the distilling flask, see figure 2. The operation is done either manually or from the software by hand or automatically through the programmable recipe. The heat flux is excellent to the glass flask, and heating is fast.

Figure 3. A schematic view of the heating mantle. (BriskHeat, 2021)

(16)

2.2.2. Distillation column

The distillation column is an essential part of the distillation system since most of the separation occurs. As the reboiler, the column type and design depend on the final application.

Generally, the fractional distillation industry uses various packing materials in the column. The packing could be bubble caps, sieve, or valve trays, or anything else for that matter; the only thing to consider is that the filling must react with the distillate, and it should withhold temperature. The purpose of this packing is to maximize the surface area on which the vapor condenses. The column designers use simulation software to determine the column dimensions and packing. Still, it is possible to do it manually with, e.g., the McCabe-Thiele method, introduced in the later chapters. In addition to dimension determination, the design verifies that the column can reach the desired purities in the chosen process. (Tham, 2016) We will conduct multiple operations on our instrument. The system is designed for a more complex application, which enabled us to neglect the design part while purchasing the system;

hence the design review is for theoretical review only.

The packed columns are called fractionating columns, and this is the option we are utilizing in this thesis. The difference with these column types is the possibility to achieve purer distillate with the fractionating column due to multiple distillations by one process. The column, which we use is a spinning band column with fifty theoretical plates. The spinning band column should achieve exceptional results in our application, which is to purify Methanol. Figure 4.

illustrates the theoretical function of fractional distillation. (Tham, 2016)

(17)

Figure 4. Theoretical fractional distillation. (Clark, 2020)

Here, the given substances A and B are distilled. If we boil at point C1, we will reach a mixture of A and B with roughly the same concentration, C2. If we were to refine that mixture further, we would reach the point of C3, where the liquid would contain more substance B.

Furthermore, with this example, we could reach even more condensed with substance B. In this process, there are two theoretical stages (reboiler is considered as one). This is the basic principle behind fractional distillation.

The mass transfer in distillation takes place at the phase interphases in the column. The reflux from the condenser inflicts material transfer between liquid and vapor, which is the critical difference between simple and countercurrent distillation. Figure 5. simplifies the material

(18)

transfer at the column stages and the reflux from the condenser back to the boiler. (Krell, 1982)

Figure 5. Simple illustration of a packed column in batch distillation. (Krell, 1982)

The mass transfer takes place here in the column as well as in the boiler. In the distilling process, the concentrations increase in both the vapor and the liquid phases. At the start of the distillation process, the first vapor, which reaches the condenser, has the same composition as the boiler. The vapor fully condensates and returns to the boiler through the column. The vapor in the column holds more of the lighter component. The net transfer is towards the liquid; as we descend the column, the net transfer shifts towards the more lightweight component, enriching the vapor with the lighter component. The material exchange continues until the system reaches equilibrium, and the heavier elements return to the boiler. (Krell, 1982)

(19)

2.2.3. Condenser

The purpose of the condenser in distillation is to condense the rising vapors back to liquid. The main difference between these condensers is the design and cooling capabilities, see Figure 6. Various condenser designs are available on the market, such as Graham, Dimroth, and Friedrichs, all named after their developer. (Krell, 1982)

Figure 6. Different condenser designs. (ChemDraw).

The critical parameters affiliated with the condenser are temperature, heat flow, and material flow. These parameters will decide whether the cooling capability is adequate for the specific process. The condenser must remove the heat flow from the rising vapors fast enough for the

(20)

vapor to cool and condense back to liquid before passing the condenser. The material flow here refers to the cooling material flow; this also affects the cooling capability. The fundamental difference between the mentioned condensers is the cooling area, which is specific to each condenser. The pragmatic view here is, the more significant the cooling surface area, the better the cooling capability, assuming other parameters remain constant.

(Krell, 1982) In addition, the coolant flows direction affects the cooling capacities. We are using downstream cooling flow, which yields us adequate cooling.

2.2.4. Reflux ratio

The purpose of the reflux flow in distillation is to return the chosen amount of substance to the distillation column to achieve better separation. The reflux is given as the ratio between the removed liquid and liquid returned to the distillation column. There are two basic operations for reflux in the batch process: constant reflux and constant composition reflux.

The difference between these operations is that the constant reflux is held at the same value throughout the separation process, whereas the constant composition reflux is varied. The constant composition reflux process requires more input in the process planning as the reflux ratio must increase steadily as the distillation continues; this explains the latter process. The overall result will be the same product-wise, but the processing time is shorter when the reflux ratio is varied. The same product quality is that the constant reflux ratio will be higher than required at the distillation start and vice versa. Figure 7. illustrates the differences between these processes, as well as the optimal process. (Gorak, Sorensen, 2014)

(21)

Figure 7. (a) constant reflux, (b). constant composition, (c). optimal reflux. (Gorak, Sorensen, 2014)

2.3.

Controlling the batch distillation process

There are numerous parameters, which affect the overall distillation process. These are, e.g., temperatures, pressures, flow rates, and different levels such as reflux ratio. All these affect one another, hence changing one parameter usually involves at least another parameter. This study is not involved in the feed's composition or the distillation equipment design, so we are left with three parameters to adjust the distillation process: process pressure, temperature, and reflux ratio. Methanol must be distilled at atm because lower pressure would cause boiling at a lower temperature. Also, the heating should remain calm to give the condenser time to cool down the vapors, so the reflux ratio is the only parameter to adjust. Since the process is not time-dependent, we use a constant reflux ratio to keep it simple. (Linninger, 2004)

We initially collected the lighter fractions in the test runs, roughly three to six degrees below the Methanol's boiling point. After the first fraction, the actual production fraction is collected using the boiling point from literature + 0.5°C – 1.0°C. (Safety data sheet, Methanol, 2021) The

(22)

equipment is software controlled, which means that it automatically collects two fractions and leaves the third in the reboiler. These settings are altered experimentally in the experiment section. The most important factors to consider during the process are the adequate cooling of the vapors, flooding of the column, pressure (if low pressures are used), and boiling. The condenser temperature should be monitored to ensure the total condensing of the vapors.

The passing vapors could harm the system if the chemical reaches, e.g., the vacuum pump, which is chemical resistant to only some extent. The flooding of the column has a direct effect on the efficiency of the column. In this system, actual flooding cannot be seen, so the reflux and the boiling are the key factors to monitor here. There is a tip on the top of the column, which can be used to check the reflux from the condenser; if the backflow is nice drizzle, the boiling and the condensing are optimum. The system has two points of temperature measurement; the boiling flask and at the top of the column. With the VLE-data for this exact system, we can determine the composition at each point when the pressure is set to constant.

(Linninger, 2004)

2.4.

Vapor-liquid equilibrium

In this thesis, we are studying the mixture of polymer synthesis waste. The assumed solution contains mainly Methanol and NMP. The boiling points for these substances are methanol 64.7°C (Safety data sheet, Methanol, 2021) and n-methyl-pyrrolidone 202°C (Safety data sheet, n-methyl-pyrrolidone, 2021) at atm. The VLE graphs for such a system are presented in the following figures. For the initial calculations, the following parameters were used:

Table 1. Antoine's parameters for chosen substances. (Dortmund data bank, 2021)

No. A B C Tmin Tmax

NMP (1) 7.54826 1979.68 222.2 6 206

Methanol (2) 7.9701 1521.23 234 65 214

Antoine Equation Parameters (P in mmHg, T in °C)

(23)

Table 2. Liquid mole fraction of Methanol. (Dortmund data bank, 2021)

The vapor pressures of this system are calculated with Antoine's equation:

𝑙𝑜𝑔₁₀(𝑝) = 𝐴 − [ ^𝐵

(𝑇+𝐶)] (1)

where p Vapor pressure of a substance T Temperature, K

P [kPa] xMeOH [mol/mol]

0.4706 0

2.6904 0.05

4.71 0.102

5.03 0.112

6.679 0.148

9.401 0.199

12.37 0.245

12.69 0.251

16.56 0.307

22.27 0.386

30.27 0.479

34.36 0.523

42.95 0.608

46.9 0.643

51.09 0.679

58.71 0.751

62.53 0.781

69.5 0.846

72.59 0.877

75.25 0.902

77.72 0.923

82.931 0.98

84.59 1

(24)

A, B, and C Substance specific Antoine's constants

Figures 8. And 9. illustrates the system in constant pressure and the temperature change effects on the composition of vapor and liquid phases in the chosen system. The figures are presented and calculated in the following graphs according to the experimental data shown in Table 2. The calculated pressures from Eq (1) are presented in table 3.

Table 3. Calculated pressures for the NMP-methanol system.

T [K] T [°C] P [kPa] xmethanol

[mol/mol] xNMP [mol/mol] Psat,methanol

[mmHg] Psat,NMP [mmHg] PTot [mmHg] ymethanol

[mol/mol] yNMP [mol/mol]

475.10 201.95 0.47 0.00 1.00 0.00 1.00 1.00 0 1

428.00 154.85 2.69 0.05 0.95 0.05 0.95 1.00 0.05 0.95

405.83 132.68 4.71 0.10 0.90 0.10 0.90 1.00 0.102 0.898

402.81 129.66 5.03 0.11 0.89 0.11 0.89 1.00 0.112 0.888

393.79 120.64 6.68 0.15 0.85 0.15 0.85 1.00 0.148 0.852

384.32 111.17 9.40 0.20 0.80 0.20 0.80 1.00 0.199 0.801

377.80 104.65 12.37 0.25 0.76 0.25 0.76 1.00 0.245 0.755

377.05 103.90 12.69 0.25 0.75 0.25 0.75 1.00 0.251 0.749

370.89 97.74 16.56 0.31 0.69 0.31 0.69 1.00 0.307 0.693

364.07 90.92 22.27 0.39 0.61 0.39 0.61 1.00 0.386 0.614

357.84 84.69 30.27 0.48 0.52 0.48 0.52 1.00 0.479 0.521

355.36 82.21 34.36 0.52 0.48 0.52 0.48 1.00 0.523 0.477

351.18 78.03 42.95 0.61 0.39 0.61 0.39 1.00 0.608 0.392

349.66 76.51 46.90 0.64 0.36 0.64 0.36 1.00 0.643 0.357

348.18 75.03 51.09 0.68 0.32 0.68 0.32 1.00 0.679 0.321

345.49 72.34 58.71 0.75 0.25 0.75 0.25 1.00 0.751 0.249

344.45 71.30 62.53 0.78 0.22 0.78 0.22 1.00 0.781 0.219

342.36 69.21 69.50 0.85 0.15 0.85 0.15 1.00 0.846 0.154

341.42 68.27 72.59 0.88 0.12 0.88 0.12 1.00 0.877 0.123

340.70 67.55 75.25 0.90 0.10 0.90 0.10 1.00 0.902 0.098

340.10 66.95 77.72 0.92 0.08 0.92 0.08 1.00 0.923 0.077

338.57 65.42 82.93 0.98 0.02 0.98 0.02 1.00 0.98 0.02

338.06 64.91 84.59 1.00 0.00 1.00 0.00 1.00 1 0

(25)

Figure 8. T-x-y plot for Methanol.

Figure 9. T-x-y plot for NMP.

(26)

For both substances, the partial pressures in liquid are calculated, as stated previously, according to Antoine's equation. We can calculate the partial pressure in vapor; here, the total pressure is set to 760 mmHg. Results are shown in figure 10.

Figure 10. NMP-methanol system.

Figures 11. and 12. show the effect of the total pressure of the NMP-methanol system on each substance, respectively.

(27)

Figure 11. Effect of total pressure in NMP-methanol system in Methanol's state of matter.

Figure 12. Effect of total pressure in NMP-methanol system in NMP's state of matter.

(28)

From the VLE data, we can construct a design for a necessary distillation column for the application – McCabe – Thiele method, which is handled more thoroughly in the next chapter.

2.4.1. McCabe – Thiele method

The design of a distillation process usually requires a rigid simulation for the whole process.

Several programs are available to conduct this task, such as HYSYS, ChemCAD, and Aspen, to name a few. This software requires detailed knowledge of energy and material balances, including dynamic equilibrium calculations making these quite complex. A few of these models are presented for illustration purposes later in this thesis. As mentioned, our instrument is predefined, so it is not feasible to do rigid simulation at this point. Hence, a simpler McCabe- Thiele -method is introduced. The technique has a few assumptions, making it easier to understand and more usable when the mentioned software is unavailable or not desired. We must make assumptions with the more straightforward method, e.g., binary mixture, constant pressure, no heat losses. These assumptions suit us, and these parameters are neglected when constructing the McCabe-Thiele -diagram. Since we have these assumptions, we have a molar overflow assumption, which means we have constant vapor and liquid streams in both sections; rectification (over the feed stage) and stripping (under the feed stage). The column balances are expressed as follows:

𝑅𝑒𝑐𝑡𝑖𝑓𝑦𝑖𝑛𝑔 𝑠𝑒𝑐𝑡𝑖𝑜𝑛: 𝑦 = ^𝐿

𝑉 𝑥 + ^𝑥^𝐷^𝐷

𝑉 (2)

𝑆𝑡𝑟𝑖𝑝𝑝𝑖𝑛𝑔 𝑠𝑒𝑐𝑡𝑖𝑜𝑛: 𝑦´ = ^𝐿′

𝑉′ 𝑥 − ^𝑥^𝐵^𝐵

𝑉′ (3)

where y mole fraction of component in the vapor, rectifying section, mol

(29)

y´ mole fraction of component in the vapor, stripping section, mol x mole fraction of component in the liquid, mol

xD mole fraction of component in the distillate, mol

xB mole fraction of component in the bottom product, mol D distillate flow, mol/h

L total flow rate in the liquid, rectifying section, mol/h L´ total flow rate in the liquid, stripping section, mol/h

V total flow rate in the vapor, rectifying section mol/h V´ total flow rate in the vapor, stripping section, mol/h

The initial phase is to calculate the operating lines for rectifying and stripping sections with the given equations. Since we do not have exact information about the system, we assumed a few parameters to conduct the calculations. These are feed composition (q), distillate (XD), and bottoms product (XB). In this graph, the reflux ratio was set to 0,5, whereas the actual experiments varied this parameter. The operating lines are in Figure 10. The equilibration data presented here were acquired from the data in the last chapter. (Górak, Sorensen, 2014)

(30)

Figure 13. Top and lower operating lines.

The desired bottom product is set to 1.0%, and the distillate purity is 99.0%. We have assumed a constant composition of 50 mol-% of the more volatile component; hence, ZF is 0.5. The next step is to determine how many stages the system would require to reach the desired result with these parameters and assumptions, see Figure 13.

(31)

Figure 14. Theoretical stages are needed to reach a purity of 99% of Methanol.

The number of ideal stages for this kind of system is 3 – 1 = 2. The first stage is always counted as the reboiler; hence it is neglected. These calculations are made with assumptions, and the example is for illustration. The analysis for the number of stages considers the stage efficiency:

the number of theoretical trays/stage efficiency. Also, the feed stage should be regarded as it depends on whether we have saturated vapor or liquid; this is neglected here because the system used in the experimental part is set to have fifty theoretical plates, and there is no possibility to change the feed line, it is the reboiler by default. (Tham, 2016)

A summary for the McCabe-Thiele method is as follows:

➢ Determine x-y diagram for the desired system, add the x = y -line

➢ Find, or assume, the bottom and distillate conditions

(32)

➢ Add the q-line

➢ Add top and lower operating lines

➢ Add step stages from top to bottom until the bottom is reached.

The McCabe-Thiele method is a pragmatic view to column design and is used here because the system used in this thesis is predesigned and not made solely for this exact process.

(Górak, Sorensen, 2014)

2.4.2. Activity models

The purpose of this chapter is to give a brief idea of what kind of models could be used in describing acquired data in the VLE process. Therefore, the chapter is expecting some level of knowledge on the subject beforehand. In this thesis, we consider only non-ideal binary mixtures and the models selected for the review, which are the most suitable for this application. The models are Wilson's equations, the NRTL model, and UNIQUAC. These models are used in the column and the distillation system design. (Elliot, J. 2012)

Wilson's equation predicts the activity coefficients in binary solutions when experimental data is not available. The lack of empirical information forces us to estimate the results from theoretical correlations. Van Laar and Margules equations are like these but require actual data to get acceptable results. The significant difference, which Wilson proposed was the concept of local composition, dealt with in the previous chapter. To calculate binary VLE data using Wilson's method, we need the following equations:

(33)

The activity coefficient for component 1:

ln 𝑦₁ = − ln(𝑥₁+ 𝑥₂Λ₁₂) + 𝑥₂( ^Λ¹²

𝑥1+ 𝑥2Λ12− ^Λ²¹

𝑥1Λ21+ 𝑥2) (4)

The activity coefficient for component 2:

ln 𝑦₂ = − ln(𝑥₁ Λ₂₁+ 𝑥₂) − 𝑥₁( ^Λ¹²

𝑥1+ 𝑥2Λ12− ^Λ²¹

𝑥1Λ21+ 𝑥2) (5)

And the calculation for the variables  and 

Λ₁₂ = ^𝑉²

𝑉1exp (^−𝐴¹²

𝑅𝑇 ) , Λ₂₁ = ^𝑉¹

𝑉2exp (^−𝐴²¹

𝑅𝑇 ) (6)

where 1 and 2 are the activity coefficients for components 1 and 2 x1 and x2 are the liquid mole fractions of components 1 and 2 V1 and V2 are the molar volumes of components 1 and 2

ij is the interaction energy between i and j R is the gas constant,

T is the temperature in Kelvin, K

(34)

The strategy here is to solve parameters 12 and 21 initially, and after, we can solve the activity coefficients of Wilson's equation. Then, we can achieve partial pressures when we apply the gained values to the modified Raoult's law. This method is usually utilized through simulation software, such as Aspen Plus. (Elliot, J., 2012)

The UNIQUAC is a continuum to Wilson's equation, making a few changes in the initial equation. The first significant difference is that the model depends on surface area rather than volumes. Wilson's equation assumes that ij = ij, whereas the UNIQUAC model expects the local composition to depend on a relative surface area of a molecule. For a binary mixture, the UNIQUAC activity models are as follows. (Elliot, J. 2012)

ln 𝛾₁ = ln^¹

𝑥1+ (1 − ^¹

𝑥1) − 5 𝑞₁ [ln^¹

θ ₁+ (1 − ^¹

θ₁)] + 𝑞₁[1 − ln(θ₁+ θ₂𝜏₂₁) −

θ₁

θ₁+ θ₂𝜏₂₁− ^θ²^𝜏¹²

θ₁𝜏₁₂ + θ₂] (7)

ln 𝛾₂ = ln^²

𝑥2+ (1 − ^²

𝑥2) − 5 𝑞₂ [ln^²

θ 2+ (1 − ^²

θ₂)] + 𝑞₂[1 − ln(θ₁𝜏₂₁+ θ₂) − ^θ¹^𝜏²¹

θ1+ θ2𝜏21−

θ₂

θ₁𝜏12 + θ₂] (8)

As seen here, the UNIQUAC model is much more complex than the Wilson model. In the activity equations, f is molecular volume fractions, and q is the surface fractions. However, there are two different parameters to adjust in both, and the Wilson equation gives almost as good predictions as UNIQUAC and is applicable at least in this thesis's practices. The significant difference here is that the UNIQUAC model is also functional in the liquid-liquid simulations, whereas the Wilson model is not applicable. (Elliot, J. 2012)

(35)

The third applicable model for VLE modeling is NRTL (Non-random two liquid). The model is also widely used in phase equilibria calculations. The model has three separate variables, which are determined by regression of experimental data. The NRTL model is applicable in VLE calculations as well as with LLE calculations. The NRTL is comparable to Wilson's equation but except that NRTL applies to liquid-liquid systems. The activity coefficient calculation with NRTL is as follows:

ln 𝛾₁ = 𝑥₂² [ ^𝜏¹²^𝐺¹²

(𝑥₁𝐺₁₂+ 𝑥₂)²+ 𝜏₂₁ ( ^𝐺²¹

x₁+ 𝑥₂𝐺₂₁)²] (9)

ln 𝛾₂ = 𝑥₁² [ ^𝜏²¹^𝐺¹²

(𝑥1+ 𝑥2𝐺21)²+ 𝜏₁₂ ( ^𝐺²¹

x₂+ 𝑥₁𝐺₂₁)²] (10)

The NRTL model neglects thermodynamic integration as there is an assumption that free energy equals excess Gibbs's energy. The parameter introduced in the final equation is a literature value with an energy unit. (Elliot, J. 2012)

When applied through the mentioned software, we get the simulated results as in Figure 15.

Here, we can conclude that the NRTL model parameters describe NMP-MeOH -system in the liquid state, but all three models are similar in the vapor state. Hence, it would be feasible to use the NRTL model on the system at hand. (Zegeibh, Takache. 2021)

(36)

Figure 15. UNIQUAC, NRTL, and Wilson's models compared to experimental data.

2.4.3. Tray efficiency and HETP

In the distillation column, the vapor and liquid are in contact. While conducting a theoretical estimation of a distillation column, such as the McCabe-Thiele method, the assumption is that the two phases are in equilibrium, and the plates function at maximum efficiency. This is not the case for most real systems because of not enough contact between phases or not enough mixing. A few different models are available, which describe the tray efficiency, from simple models to more complex ones. Height equivalent to theoretical plates (HETP) is used to consider the height of the packing, which could achieve the same separation as theoretical plates. The method is mainly in use with packed columns. As we consider the tray column, it is more feasible to estimate the plate efficiency by, e.g., Murphree's vapor efficiency since it does not consider what kind of packing is used. The Murphree's vapor efficiency is as follows:

(37)

𝐸_𝑀𝑉 = ^𝑦^𝑛,𝑖^{− 𝑦}^{𝑛−1,𝑖}

𝑦_𝑛,𝑖^∗ − 𝑦_{𝑛−1,𝑖} (11)

The composition of the vapor phase determines the vapor efficiency as it reaches a plate and leaves the plate. The term, 𝑦_𝑛,𝑖 – 𝑦_{𝑛−1,𝑖}, is the actual change of composition in each tray, and term 𝑦_𝑛,𝑖^∗ − 𝑦_{𝑛−1,𝑖}, is the change in equilibrium case. Murphree's method assumes a perfect mixing on each tray. This model would be a valid model to calculate the actual tray quantity needed for this separation, but it is not necessary because of the setup used in this thesis.

(Burns, 1969)

Murphree's method is a popular choice for estimating tray efficiency. It is closer to actual cases than overall efficiency calculation but more straightforward to use than a more precise local efficiency method. (Geankoplis, 2013)

(38)

3. Commercial distilling devices

Before this thesis, a batch distillation system had to be acquired to conduct the empirical part of the study. There are numerous commercial systems readily available on the market. The starting point for us was the final application. From this initial fact, we knew that we had to search for a fractional distilling device. Here is an overview of the considered options.

3.1.

BR Instruments

BR Instruments has a variety of solutions for distilling applications. The spinning band column system was our primary focus. The system has a boiler, spinning band column, reflux, and a condenser. In addition, there is a vacuum system and a chiller as an option. The spinning band column in the thermodynamical sense has fifty theoretical plates. The spinning band is a Teflon helix located inside the column. Temperature is measured at the pot and the top of the column. The system is equipped with software, enabling automatic control of the process from start to finish. This kind of system is well suited to our needs, as reached purities are >99%, and ease of use. The system is equipped with two fraction collectors by default, and it is expandable to four collectors. An illustration of the system is in Figure 16, excluding the vacuum pump. (Br instruments, 2021)

(39)

Figure 16. The Br Instrument's fractional distilling device. (Br Instruments, 2021)

3.1.1. BR Instruments software

Automation was one key component when deciding since the distillation would be relatively long. The chosen instrument has software, which controls the overall setup. This is a convenient way to use the system, and because of the software, the whole system is semi- automatic. This will enable the system to work also outside business hours. The software's front page is shown in Figure 17. (Br Instruments, 2021)

(40)

Figure 17. Start screen of BR9600. (Br Instruments, 2021)

From the "Manual controls," the user can manually control the system, see Figure 18.

(41)

Figure 18. Manual controls of the BR9600 instrument. (Br Instrument, 2021)

In the previous Figure, there is shown the parameters, which could be controlled manually.

These are Condenser bath temperature, Band motor, Reflux valve, and vacuum level. The range for the vacuum is 0 mmHg to 100 mmHg, so the window is relatively narrow. The

"Manage Run Data" -menu lets the user check, change, or create new recipes. The following Figure shows an example recipe for methanol distillation. (Br Instruments, 2021)

(42)

Figure 19. The recipe window for Br 9600 instrument. (Br Instruments, 2021)

Here, the user can define the desired distillation recipe. First, the user must specify the run pressure, heating, equilibration time, number of cuts, and maximum pot temperature. The parameters are explained next.

Table 4. the parameters from the previous Figure.

1. Run pressure Overall pressure where the distillation is conducted 2. Initial heat Heat in the beginning

3. Equilibration Equilibration time before 1^st cut is taken out

4. Number of cuts A user-defined nr. of cuts. We have a maximum of two cuts, limited by the receiver's quantity. Up to eight would be possible

5. Max pot temperature Maximum temperature for the distilling pot will end the process if reached

6. Cut number Number of the cut, which is in process 7. Open cut The temperature where the cut is taken out

8. Close cut The temperature where the cut is finished, if the second one, this will end the process

9. Reflux ratio The reflux ratio - explained in the previous chapters 10. Heat rate The heat rate which heats the distilling pot

11. Condenser Temp. Condenser temperature

(43)

The last menu will let the user start the distillation, "Start distillation." After choosing this option, the software forces the user to select the appropriate recipe and check that the instrument is ready for the start.

Figure 20. Start of the distillation. (Br Instruments, 2021)

After the start of the distillation, the user can see the logged data as in Figure 21.

(44)

Figure 21. The logged data screen. (Br Instruments, 2021)

In addition to the previous, the software allows the user to see which step is underway, as in Figure 22.

(45)

Figure 22. Software show which steps are done. (Br Instruments, 2021)

Figure 19. is the same as shown previously, but here, the run has ended. The dark grey color indicates which step is done. The data output, which the software records, are as in Figure 21.

(46)

Figure 23. Methanol distillation with BR9600.

Here, we can see a few of the parameters the system will log. The data is from the methanol distillation process. The process starts from scratch and will heat the pot to vaporize the Methanol. The equilibration time is set to four hours to maximize the purity of our product.

The first cut is collected after the equilibration. We can conclude from the data that there is a relatively small quantity of low boilers, as the system will start the methanol distillation almost straight after that time, see Figure 21. The change between cuts is so fast that the logging period cannot record the change. The overall process time was roughly twelve hours, and the process stopped after the vapor temperature reached the given parameter, 65.5°C.

(47)

Figure 24. The exact Figure is the nr. 20, but the time scale is shortened to illustrate the change between cuts.

In the data, there are other parameters also, which indicate the different phases of the process. These are the run mode and cut. The run mode indicates the stage the system is in at the given moment, and the cut tells the users which cut is under distilling at which point.

(Br Instruments, 2021)

3.2.

Goel Scientific

The second option was Goel Scientific. They offer a fractional distilling device with roughly the same setup as the previous one. However, the total system is somewhat more complicated

(48)

and more laboratory-like compared to the last. The solvent recovery system's schema is presented in Figure 25.

Figure 25. Schematic illustration of Goel's instrumentation. (Goel Scientific, 2014)

As stated, the overall system is the same with both, but the Goel system has more piping and glassware. The column packing or the packing material was not made aware of at the time of inquiries. The weakness here is the lack of software and software drivability, making this a non-standalone system. There is also the possibility of collecting two fractions in this system because of the piping and valve system. (Goel Scientific, 2014)

(49)

3.3.

Sigma scientific glass

The third option was Sigma Scientific, which has various options for distilling processes. The setup is comparable to the latter one. The same options are available, like a vacuum and different heating mantles. The column here is a fractionating column, which is the only information we have about this one. The difference here is the one fraction collection by default; see Figure 26. The equipment itself would still be valid for our application. There are still the same features lacking, which the latter system lacks too. The BR Instruments has the software advantage over their Indian counterparts. The software is a considerable improvement to the process as the software is programmable for a specific purpose. (Sigma Scientific Glass, 2021)

Figure 26. The Sigma Scientific's fractional distillation system. (Sigma Scientific Glass, 2021)

(50)

3.4.

Comparison of commercial distillation systems

These were the options, including building the system ourselves. The overall solution, software, and the response to our application needs were the conclusive factors in this decision-making. Technical specifications are presented in Table 5.

Table 5. A comparison of the technical data of the systems.

Column Distillation flask

Condenser Heater Condensing bath

Reflux Fraction collector Br

Instruments

Teflon spinning band.

22 dm³ Inner coil + outer jacket

Kettle -20 °C – 140

°C

Automatic Up to 8 fractions

Goel Scientific

Packed column

20 dm³ A = 0.35 m²

Kettle Optional Manual Manual

Sigma Glass Packed column

20 dm³ A = 0.35 m²

Jacketed Optional Manual Manual

All systems also had an optional vacuum system, which was essential as we wanted low- pressure distillation. From a technical point of view, the Br Instrument stands out with a more automatic system than its rivals. The two latter ones have similar systems in parameters. From a stand-alone point of view, we decided to purchase the Br Instruments system.

(51)

4. Experimental section

The experimental study took place at the polymer laboratory of the company in question.

Moreover, the material was acquired from the actual polymer synthesis process from the precipitation and washing processes. These processes happen in various steps, and the material was collected from arbitrary phases. Therefore, the material is suspected of containing Methanol, NMP, and different other unknown salts and chemicals. The goal was to acquire as pure solvent as possible with the fractional distilling device. The overall distill was analyzed by a third-party operator since we do not have the facilities available for analysis.

4.1.

Solvents

Methanol is toxic alcohol, which is at atm a colorless liquid and withholds a distinct odor. The chemical is widely used, e.g., as a solvent or anti-freeze substance. The toxicity is caused by ingesting the substance as the human body will oxidize the chemical to formaldehyde and carbon dioxide. NMP, on the other hand, is not toxic, although it has been illustrated to affect the reproductive system in large or long exposures. NMP is also a widely known and used solvent in the drug and polymer industry. NMP is also set on a watch list by the ECHA, which is the main reason we want to clean the solvent and decrease the overall consumption. (Safety data sheet, Methanol. 2021, Safety data sheet, n-methyl-pyrrolidone, 2021)

4.2.

Test distillations

The samples are collected from the different process stages, so the solvents should be as dirty as possible, a worst-case scenario. As mentioned previously, the total content of the waste is unknown and not interesting in this sense. We know that most waste is Methanol; hence, the

(52)

test phase purifies the methanol fraction. The test material was collected in a joint container to get the maximum dissolved material in the distilling pot. The material collection happened from different batches and different stages of the process. Four separate process tests were conducted. All processes used different settings for the parameters up for variation; see Table 6. Constant parameters were:

1. Run pressure, 760 mmHg 2. Number of cuts, 2

3. Condenser temperature, 0°C

Table 6. Test run parameters

1^st test 2^nd test 3^rd test 4^th test

Equilibration time 10 min 30 min 60 min 240 min

Max. pot temperature 66°C 75°C 75°C 80°C

Open 1^stcut 60°C 60°C 62°C 62°C

Close 1^st cut 64°C 64°C 64°C 64°C

Reflux ratio, 1^st cut 5:1 2:1 2:1 2:1

Open 2^nd cut 64°C 64°C 64°C 64°C

Close 2^nd cut 65°C 65°C 64.5°C 64.5°C

Reflux ratio, 2^nd cut 5:1 3:1 5:1 5:1

The test sequence was constructed with the aid of suppliers' application notes and according to best guess. The equilibration time is between the start of the boiling and before the 1^st cut is taken out. The cuts are collected automatically by the fraction collector. The overall purpose was to iterate the process parameters to reach the desired goal – to optimize the process

(53)

according to purity. In the test runs, the goal was to achieve desired purity for Methanol (~99%). The results are evaluated in chapter 5.

4.3.

Analysis methods

The test run products were analyzed by two different methods: Gas chromatography and Karl- Fischer titration. The prior one is used in this case to study the content of the sample, and the latter is to determine the residue water content of the samples. The analysis was conducted in collaboration with a third party, as we do not have this capability in-house.

4.3.1. Gas chromatography

The purified samples were sent to Gas chromatography analysis; moreover, HS-GC-MS (Headspace gas chromatography with mass spectrometer). The method is suitable for the quantification of the selected solvent. The sample is vaporized into the GC's column in the GC analysis, and the retention time is measured. The overall system is composed of Figure 27.

Figure 27. Schematic illustration of HS-GC-MS. (Emvas et al., 2015)

(54)

The carrier gas used is usually nitrogen due to its inert nature. In the headspace oven, the samples are put into a vacuum and vaporized. The sample is taken from the gas phase and sucked into the GC oven, where the actual column is placed. The column is equipped with a suitable stationary phase. Different molecules from the headspace reach the detector at various times; hence it is possible to determine the sample's composition. The retention time is affected by the molecule's molecular weight and polarity of the molecule. From this information, we can determine the boiling point of the molecule and its composition. The concentration is evaluated with an internal standard. This standard is specific for different types of molecules, and the choice of the standard is crucial for the analysis. If we want to quantify the sample by looking at the peak area, the response factor must be determined. The response factor method is the most common way. The response factor tends to vary from system to system. Because of the variation could be more feasible to use the relative response factor and internal standard mentioned earlier. This method can be used to determine an unknown concentration with a known concentration. (Bouchonnet, 2013) Here, further examination is neglected because the analysis is not in the scope of this thesis.

4.3.2. Karl-Fischer titration

The Karl-Fischer titration (later KFT) is a widely used water content analysis method. The KFT is applied in multiple disciplines every day, from the food industry to the pulp and chemical industries. In this thesis, the KFT is used alongside gas chromatography to get a broader view of the contents of our sample, although the water content should not perturb the solvent reuse in our application. The KFT is divided into two different sub-categories: volumetric and coulometric. The process differs between these two, and the major decisive issue is the expected water content. We expect to have relatively high water content, so the volumetric

(55)

analysis suits us. For smaller concentrations, below 1 mg/dm^3,the coulometric is a better option. (Rouessac, et al., 2007)

The volumetric KFT is started by measuring the water content of the KFT reagent for a baseline, respectively. The humidity of the surroundings will interfere with the measurement;

hence the system must be isolated from the environment. The overall setup is as follows:

Figure 28. KFT instrumentation. (Rouessac, et al., 2007)

The KFT cell contains the sample, nitrogen, electrodes, and two different reagents. The overall purpose is to titrate a reagent to the sample, which reacts with the water to achieve a solution that does not conduct electricity. The amount of reagent added will determine the water content. In other words, we measure the amount of reagent used to neutralize the water in the sample. (Rouessac, et al., 2007)

(56)

4.4.

Test samples

The results were analyzed according to the methods mentioned in the previous chapter—the initial screening with GC-MS and a composition analysis with FID -detector. In addition, the water content was analyzed with KFT. The overall test batch quantity was four. All distillations were conducted in different settings. The settings were mentioned in chapter 4.2. The results were compared with a pure sample, which acted as a reference measurement. The reference was obtained straight from the manufacturer's flask and is assumed to be pure Methanol.

Below is a matrix of the test batches.

Table 7. Matrix of the samples and performed measurements.