Process Optimization of Ultrasonic Spray Coating of Polymer Films

Subscriber access provided by HARVARD UNIVERSITY
Article
Process Optimization of Ultrasonic Spray Coating of Polymer Films
Sanjukta Bose, Stephan S. Keller, Tommy S. Alstrøm, Anja Boisen, and Kristoffer Almdal
Langmuir, Just Accepted Manuscript • DOI: 10.1021/la4010246 • Publication Date (Web): 30 Apr 2013
Downloaded from http://pubs.acs.org on May 14, 2013
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Langmuir is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.
Page 1 of 30 Langmuir
1
2
3
4
5
6
Process Optimization of Ultrasonic Spray Coating of
7
8 Polymer Films
9
10 Sanjukta Bose1, Stephan S. Keller*1, Tommy S. Alstrøm2, Anja Boisen1 and Kristoffer Almdal1
11
12
13 1
14 Department of Micro- and Nanotechnology, Technical University of Denmark, DK-2800 Kongens
15
16 Lyngby, Denmark
17
18
19 2
20
Department of Applied Mathematics and Computer Science, Technical University of Denmark, DK-
21
22 2800 Kongens Lyngby, Denmark
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
Langmuir Page 2 of 30
1
2
3
4
5 ABSTRACT
6
7 In this work we have done a detailed study of the influence of various parameters on spray coating of
8
9 polymer films. Our aim is to produce polymer films of uniform thickness (500nm-1µm) and low
10
11
12
roughness compared to the film thickness. The coatings are characterized with respect to thickness,
13
14 roughness (profilometer) and morphology (optical microscopy).
15
16 Polyvinylpyrrolidone (PVP) is used to do a full factorial design of experiments with selected process
17
18
19 parameters such as temperature, distance between spraying nozzle and substrate and speed of the
20
21 spraying nozzle. A mathematical model is developed for statistical analysis which identifies the
22
23 distance between nozzle and substrate as the most significant parameter. Depending on the drying of
24
25
26 the sprayed droplets on the substrate we define two broad regimes, ‘dry’ and ‘wet’. The optimum
27
28 condition of spraying lies in a narrow window between these two regimes where we obtain a film of
29
30
desired quality. Both with increasing nozzle-substrate distance and temperature the deposition moves
31
32
33 from a wet state to a dry regime. Similar results are also achieved for solvents with low boiling points.
34
35 Finally, we study film formation during spray coating with poly (D, L-lactide) (PDLLA). The results
36
37
38
confirm the processing knowledge obtained with PVP and indicate that the observed trends are
39
40 identical for spraying of other polymer films.
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58 2
59
60
1
2
3
4
5 INTRODUCTION
6
7 Polymer thin film coatings are a subject of interest for a broad scientific community due to a large
8
9 number of potential applications. Their use ranges from photovoltaic cells1, electrochemical devices2,
10
11
12
chemical3 and biological sensors4 to biomedical applications like drug delivery, tissue engineering and
13
14 biocompatible implants5. Among a host of possible polymer film coating techniques such as spin
15
16 coating, bar coating, inkjet printing and screen printing6, spray coating has come up with the advantage
17
18
19 of being able to coat three dimensional microstructures like cantilevers7 ,micro-needle arrays for drug
20
21 delivery8, and non-planar surfaces for nanoelectromechanical systems (NEMS) and
22
23 microelectromechanical systems (MEMS) applications9,10. Pham et al. have shown the advantage of
24
25
26 spray coating photo resist on high topography surfaces for applications in MEMS11. Pan et al. have
27
28 used spray coating to fabricate drug eluting stainless steel stents by coating them with biodegradable
29
30
polymer acting as drug carrier12.
31
32
33 Spray coating in general includes the formation of fine droplets of a coating solution and their
34
35 deposition and coalescence on a substrate. The liquid to be used for coating breaks up into small
36
37
38
droplets by atomization in the spraying nozzle. The generated droplets are then deposited on a substrate
39
40 kept below the spraying nozzle where they coalesce to form a coating. Conventionally, air pressure is
41
42 used for the atomization where the liquid coming through the nozzle is broken down into tiny droplets
43
44
45 upon passage through a narrow orifice13. Alternatively the ultrasonic atomization technique where high
46
47 frequency sound vibration generates a fine mist of solution is used for spray coating and often preferred
48
49 over pneumatic atomization14. Overall, the advantages of ultrasonic spray coating are: simple, good
50
51
52 transfer efficiency, economical, precise conformal coating on planar and non-planar surfaces,
53
54 generation of small droplets in the micrometer range and good reproducibility11,15. Spray coating is
55
56
associated with many material and process parameters which must be controlled in order to obtain a
57
58 3
59
60
1
2
3
4
5 uniform film. The concentration of the polymer solution, the solvent used, the substrate on which the
6
7 solution is sprayed, the number of spray passes, the flow rate of the polymer solution, the temperature
8
9 of the substrate, the distance between the spray nozzle and the substrate and the speed of the moving
10
11
12 nozzle during spraying along with the air pressure which transports the sprayed droplets to the substrate
13
14 are all to be taken into consideration. The large numbers of parameters indicate the need for systematic
15
16 process optimization. Perfetti and co-workers16 and McGrath et al8 have investigated the influence of
17
18
19 different spraying parameters based on design of experiments (DoE) and subsequent experimental
20
21 analysis. Perfetti reported the distance between spray nozzle and substrate as one of the main factor
22
23
24
affecting film uniformity. McGrath highlighted the physicochemical properties of coating solution and
25
26 substrate composition as one of the major factors determining the coalescence of droplets after
27
28 deposition.
29
30
31 Here we report a detailed study of the influence and inter-correlation of various parameters on the
32
33 quality of the resulting coating. An ultrasonic spray nozzle with programmable x-y motion is used for
34
35 spraying on a substrate placed on a stationary temperature controlled stage. We have used an aqueous
36
37
38 solution of the biocompatible polymer polyvinylpyrollidone (PVP) to do a full factorial design of
39
40 experiments with selected process parameters such as temperature, distance between spraying nozzle
41
42 and substrate and speed of the spraying nozzle. The statistical analysis is used as a guideline to perform
43
44
45 a more in depth study of the most influencing parameters. Our observations lead to the conclusion that
46
47 the optimum parameter space for spray coating is narrow. To produce a smooth film one needs to find
48
49
50
conditions which are on the border between a dry regime – where dried sprayed drops cover the
51
52 substrate surface and a wet regime – which is equivalent to casting of a solution on the surface
53
54 followed by slow drying. We use a shadow mask (described in detail in the experimental section) to
55
56
57
58 4
59
60
1
2
3
4
5 prevent some area of the substrate from coating with the spraying solution. In preliminary experiments,
6
7 the well-known “coffee ring effect” has been observed which is described by Deegan et al. as the result
8
9 of capillary flow17 where liquid evaporates faster from the pinned contact line of a deposited solution
10
11
12 and is replenished by additional liquid from inside. Accumulation of polymer after solvent evaporation
13
14 is seen at the edges of the coated areas defined by the shadow mask as also observed by Dubourg et
15
16 al.18 We have named these structures ‘edge peaks’. The aim of our study is to eradicate this edge effect
17
18
19 and to have a continuous film of PVP with uniform thickness (500nm-1µm) and minimal roughness
20
21 compared to the film thickness. In a final study we utilize and generalize our gained understanding to
22
23
24
spray coat a solution of the biodegradable polymer poly (D, L-lactide) (PDLLA) in cyclohexanone.
25
26
27
28 EXPERIMENTAL SECTION
29
30
Materials
31
32
33 PVP (K30, Mw 40,000g/mol) and PDLLA (ester terminated, Mw 16,000g/mol) were obtained in powder
34
35 form. Solvents used were cyclohexanone [boiling point (b. p). 155°C, density (ρ) 0.947g/ml at 20°C],
36
37 dichloromethane (b. p. 40°C, ρ 1.33g/ml at 20°C), toluene (b. p. 110°C, ρ 0.86g/ml at 20°C) and water.
38
39
40 All chemicals were procured from Sigma Aldrich and used as received.
41
42 Process optimization was done with solutions of PVP in water and PDLLA in cyclohexanone.
43
44
45
Solutions of PVP in dichloromethane and PDLLA in dichloromethane and in toluene were used to
46
47 compare the effect of the variation of solvent boiling point on film formation. For all experiments,
48
49 0.5wt% of polymer solution was prepared by stirring at room temperature.
50
51
52
53
54
55
56
57
58 5
59
60
1
2
3
4
5 Spray coating
6
7 Spray coating was done in an Exacta Coat Ultrasonic Spraying System (Sonotek, USA) equipped with
8
9 an AccuMist nozzle. The main components are illustrated in Figure 1. The polymer solution was placed
10
11 in a syringe pump connected by tubing to the atomizing nozzle. The pump was programmed to a
12
13
14 constant liquid infusion rate of 0.1ml/min. The tip of the ultrasonic atomizer nozzle was actuated at a
15
16 frequency of 120 kHz with a generator power of 1.3W. Compressed nitrogen at 0.03bar was
17
18
introduced into the air focusing shroud. The gas flow and the position of the shroud control the width
19
20
21 and shape of the spray coating beam. The movement of the nozzle was controlled by an x-y-z stage and
22
23 the nozzle path is shown in Figure 2. The distance between two nozzle paths was kept at 2mm which is
24
25
26
smaller than the width of the spray (~4mm). The substrate to be coated was placed on the hot plate and
27
28 the temperature was varied according to the experimental design. Microscope glass slides were used as
29
30 substrates unless otherwise stated. The number of spray passes was kept constant at 10 for each sample
31
32
33 in the study.
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56 Figure 1. Schematic representation of setup for spray coating
57
58 6
59
60
1
2
3
4
5
6
7 Characterization
8
9 A shadow mask (Figure 2a) was made by milling holes of diameter 5mm and length 20mm in a
10
11 76mm×26mm×4mm aluminium block. This mask was placed on the glass slide before deposition to
12
13
14 prevent spraying on some of the area. The masked areas acted as a baseline for characterizing the final
15
16 coating by a surface profilometer Dektak8 (Veeco, USA) from which the thickness (t in nm) and the
17
18
average roughness (Ra in nm) value were calculated as shown schematically Figure 2b. Ra also known
19
20
21 as the centre line average (CLA) is the arithmetic average deviation from the mean line over one
22
23 19
sampling length . Stylus tip radius of the profilometer was 5µm and 3mg force was applied during
24
25
26
scanning with a scan resolution of 0.28µm. The stylus was moved 10mm across the sample including
27
28 masked and deposited part. Within the 5mm wide deposited section defined by the mask window
29
30 (Figure 2a), 2mm in the centre was used as reference for calculating t and Ra to avoid influence of edge
31
32
33 effects on these values. Each sample was scanned three times and the average of the data is reported.
34
35 The surface texture was observed with an optical microscope (OM) (Zeiss, Germany) in bright field
36
37 mode.
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58 7
59
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25 Figure 2. (a) Shadow mask machined from an aluminium block placed on a microscope slide (b)
26
27 Schematic showing shadow mask effect on the sprayed polymer with profilometer scan for measuring t
28 and Ra of deposited film and the spray path of the moving nozzle with a separation of 2mm
29
30 Design of experiments
31
32 Before this optimization, several screening experiments were performed to understand the role of
33
34
35
different parameters on the film formation. The parameters tested were liquid infusion rate, generator
36
37 power, nitrogen pressure in the air focusing shroud, concentration of the polymer solution, temperature
38
39 of the substrate, the distance between nozzle and substrate and the speed of the spraying nozzle. The
40
41
42 Infusion rate was kept constant at 0.1 ml/min. At higher infusion rate polymer was seeping below the
43
44 shadow mask and the pattern required for further characterization was hampered. For the generator
45
46 power, a stable spray of the material was obtained at 1.3 W. A lower power did not form a fine mist of
47
48
49 spray due to insufficient atomization and a higher power can rip off the liquid causing large chunks of
50
51 material to be expelled, rather than a spray of droplets. The compressed nitrogen pressure in the air
52
53
focusing shroud was fixed at 0.3bar. It was found that a pressure below 0.3bar was insufficient to focus
54
55
56 the entire material coming out of the nozzle onto the substrate. A pressure higher than 0.3bar was
57
58 8
59
60
1
2
3
4
5 putting an excess thrust on the focused spray beam reaching the substrate which led to material
6
7 deposition below the shadow mask. Experiments were conducted with 0.1wt%, 0.5wt% and 1wt%
8
9 concentration of polymer solution. 0.1wt% solution had a very low solid content which was not
10
11
12 sufficient for forming a film. Furthermore, the dilute solution was seeping below the shadow mask.
13
14 1wt% polymer solution formed a film with thickness higher than the desired thickness. Therefore, all
15
16 the experiments were performed keeping the concentration of the polymer solution constant at
17
18
19 0.5wt%.Based on the screening experiments, it was concluded that the temperature of the substrate (T),
20
21 the distance between nozzle and substrate (N) and the speed of the spraying nozzle (S) were the most
22
23
24
important operational variables influencing the uniformity of the final film. These three critical process
25
26 parameters were thus used as factors for DoE. Three values of each variable were selected with T
27
28 (20°C, 40°C and 60°C), N (70mm, 50mm and 30mm) and S (10mm/s, 30mm/s and 50mm/s) for a 33
29
30
31 full factorial design. Three centre point replicates were performed to improve model adequacy.
32
33 In total 30 experimental runs were performed in random order. MATLAB R2012b software package
34
35 (Math Work Inc., USA) was used to develop mathematical models to determine the influence of the
36
37
38 design factors. The responses measured were appearance of edge peaks, thickness and roughness of the
39
40 sprayed films. The appearance of edge peaks was determined by a support vector machine (SVM) with
41
42 a linear kernel 20
as we wanted to categorize our observations in two regions; one with and the other
43
44
45 without the presence of peaks. SVM is a maximum margin method, so the decision line that is
46
47 identified by SVM will be placed with equal distance to the peak and non-peak points that define the
48
49
50
support vectors. To improve the accuracy of the model for the other two responses, roughness values
51
52 higher than 1000nm were not included in the analysis. These points were identified as outliers. The
53
54 statistically significant variables, interactions and second orders on the response roughness and
55
56
57
58 9
59
60
1
2
3
4
5 thickness were found. Formally, the following regression model is applied, both to model the
6
7 roughness and the thickness;
8
9
10 Y = β + α1T + α 2 S + α 3 N + α 4TS + α 5TN + α 6T 2 + α 7 S 2 + α8 N 2 (1)
11
12 Where Y is either the roughness or the thickness, T is the temperature, S is the speed of spraying nozzle
13
14
and N is the nozzle-substrate distance. In order to test if a factor is significant, the associated regression
15
16
17 coefficient is tested for being different than zero, i.e. if α1 is significantly different than zero then the
18
19 temperature is a significant factor. Hence, the following null hypothesis is employed
20
21
22 H 0 : αi = 0
(2)
23 H1 : α i ≠ 0
24
25
26 The hypothesis testing is carried out using the t-test.
27
28 The model specified in equation (1) is fitted using a stepwise approach. Here, the model is created
29
30
using selection algorithms in order to identify the simplest possible model to explain the observations.
31
32
33 The algorithm starts with the full model, and then iteratively removes terms that are found
34
35 insignificant, always choosing the most insignificant term. The term is identified using the F-test with a
36
37
38
threshold chosen at P =0.1. The algorithm is implemented in MATLAB R2012b21. From the results
39
40 obtained from the analysis of the model, a second set of experiments was done where the nozzle
41
42 substrate distance was changed in small steps at a fixed speed of spray and temperature. The results
43
44
45 from this second set of experiments were not included in the definition of the model.
46
47 RESULTS AND DISCUSSION
48
49 Spray Coating with PVP
50
51
52
Response edge peaks
53
54
55
56
57
58 10
59
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24 Figure 3: Plot of the decision line separating the peak area with the non-peak area. Schematics of the
25
26 deposition profiles in both areas are shown.
27
28 First the appearance of edge peaks is analyzed. Figure 3 shows the decision line predicted by the SVM
29
30 analysis. It clearly categorizes the whole experimental parameter space into two regions of a peak area
31
32
33 and a non-peak area. As seen from Figure 3, both the temperature of the substrate and the nozzle-
34
35 substrate distance play influential role in determining the appearance of edge peaks whereas the speed
36
37 of spray has no influence. When the sprayed film dries slowly , accumulation of polymer after solvent
38
39
40 evaporation due to the “coffee ring” effect17 gives rise to the edge peaks. These peaks are believed to
41
42 form due to the mobility of the polymer in the wet state. The evaporation of solution is higher at the
43
44
edge of the coated area which results in movement of polymer solution from the center to the edge.. We
45
46
47 define a condition like this as ‘wet regime’ corresponding to the peak area in Figure 3. This
48
49 phenomenon of edge peak formation has been reported both in case of spray coating18,22 and inkjet
50
51
52 printing23. In the “dry regime” due to rapid drying of the sprayed droplets peak formation is not
53
54 observed and hence that corresponds to the non-peak area in Figure 3. An increase in both temperature
55
56 and nozzle-substrate distance can pull the system towards the non-peak region. Increasing the
57
58 11
59
60
1
2
3
4
5 temperature of the substrate induces faster drying of the sprayed droplets on the substrate while an
6
7 increase in nozzle-substrate distance results in drier droplets before landing on the substrate. Fine
8
9 tuning of temperature and nozzle-substrate distance is done in the dry regime or non-peak area to reach
10
11
12 the optimum point in terms of desired thickness with low roughness.
13
14 Response thickness
15
16 The P-values predict the terms with most statistical significance, the lower the value the higher is the
17
18
19 influence of that factor on that particular response24. Here, the factors with P-values lower than 0.1 are
20
21 considered for the model. For the response thickness, P-values for the factors included in the model are
22
23
24
shown in Table 1.
25
26
27
28
29 Table 1. Factors included in the model and their corresponding P-values for predicting response
30
31
thickness.
32
33
34
35 Factors T S2 S T×S T2
36
37 P-value 0.010 0.013 0.015 0.054 0.068
38
39
40
41
42 Table 1 shows that for response thickness, temperature and speed of spray are the most significant
43
44
factors statistically followed by their interactions. The distance between nozzle and substrate plays no
45
46
47 role in determining the thickness of the sprayed film. This is reasonable as the amount of material
48
49 sprayed remains unchanged with a change in nozzle-substrate distance.
50
51
52
53
54
55
56
57
58 12
59
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24 Figure 4: Plot of thickness dependence on temperature at different speed of spraying nozzle predicted
25
26 by the model
27
28 Figure 4 shows the thickness of the sprayed film as predicted by the model at different speeds of spray
29
30 and temperature. From Figure 4 it is seen that as the speed of the spraying nozzle decreases the
31
32
33 thickness increases due to a higher amount of polymer sprayed on the substrate However, it is also
34
35 observed from Figure 4 that the thickness increases with an increase in temperature. This effect is most
36
37 prominent for the slowest speed of 10mm/s. In principle, an identical amount of polymer should be
38
39
40 deposited independent of temperature. A possible explanation of this observation is given when
41
42 considering the profilometer scans of two depositions at speed 10mm/s and nozzle substrate distance
43
44
50mm at 20°C (Figure 5a) and at 40°C (Figure 5b). It is observed from Figure 5a that the deposition is
45
46
47 more uniformly distributed over the whole surface at 20°C compared to the one at 40°C (Figure 5b)
48
49 where the surface becomes rough and the deposition builds up in the middle. At high T, the polymer
50
51
52
droplets dry quickly after they land on the substrate. The profile follows a Gaussian distribution due to
53
54 deposition of polymer on the shadow mask near the edge. At low T, the droplets dry slowly and the
55
56
57
58 13
59
60
1
2
3
4
5 polymer has mobility for long time. This potentially allows the droplets to coalesce and form dense
6
7 uniform film of low roughness and thickness compared to the one seen at high T.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Figure 5. Profilometer scan at (a) 20°C and (b) 40°C at nozzle-substrate distance 50mm and speed of
34 spray 10mm/s
35
36 Response roughness
37
38
39 For response roughness, the P-values for the factors included in the model are shown in Table 2.
40
41
42
43
Table2. Factors included in the model and their corresponding P-values for predicting response
44
45 roughness.
46
47
48 Factors N T S×N T×S T2 S
49
50 P-value 0.00029 0.00048 0.0007 0.008 0.07 0.2456
51
52
53
54
In Table 2, it is seen that the distance between nozzle and substrate is the most significant parameter
55
56 influencing film roughness, as also found by Perfetti et.al16, followed by temperature and the
57
58 14
59
60
1
2
3
4
5 interaction between S and N and between T and S. Interestingly, the S is t only found statistically
6
7 significant as an interaction term not as a single parameter. Although S shows a P-value higher than the
8
9 threshold value of P defined for the model, it was included in the model due to the interaction terms.
10
11
12 The roughness contours (Figure 6) with the edge peak area from Figure 3 (shaded region) are plotted
13
14 with T and N as variables with 10mm/s (Figure 6a), 30mm/s (Figure 6b) and 50mm/s (Figure 6c). All
15
16 the contours form curved lines due to the S×N interactions.
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58 15
59
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52 Figure 6. Contour plot of the roughness at speed of spray (a) 10mm/s, (b) 30mm/s and (c) 50mm/s as
53
54 predicted by the model. The grey region marks the peak area from Figure 3.
55
56
57
58 16
59
60
1
2
3
4
5 The roughness contours in Figure 6 show that with increase in temperature the films are getting rougher
6
7 irrespective of the speed of spray.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26 Figure 7: OM images showing the gradual increase in roughness of the spray coated film at a speed of
27 10mm/s and 45mm nozzle-substrate separation at (a) 20°C, (b) 30°C, (c) 40°C and (d) 60°C
28
29 Figure 7 shows the increase of surface roughness of the sprayed film with increase of temperature of
30
31
32
the substrate. In Figure 7 it is observed that as the temperature of the substrate is raised above 20°C
33
34 (Figure 7a), the deposition is moving from the wet regime to the dry regime. Above 30°C the surface
35
36 morphology shows the appearance of dried droplets on the film surface (Figure 7b). The amount of
37
38
39 dried droplets increases with an increase in temperature. At 60°C (Figure 7d), the whole surface is
40
41 covered with dried droplets. The roughness of the film is seen to depend on the drying of the sprayed
42
43 droplets on the substrate. Thus the most suitable temperature for spray coating is 20°C.
44
45
46
47 Considering the contours as estimated by the model, roughness is lowest in the non-peak region for
48
49 speed of spray of 50mm/s (Figure 6c). But at this highest speed of spray it is observed that the film
50
51 formation is not complete as the amount of deposited material is not sufficient to coalesce and to form a
52
53
54 continuous film.
55
56
57
58 17
59
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22 Figure 8. Profilometer scan at speed of spray (a) 10mm/s, (b) 30mm/s and (c) 50mm/s at nozzle-
23
24 substrate distance 70mm at 20°C.
25
26 This observation is confirmed by Figure 8 which shows a comparison between three profilometer scans
27
28 over identical same scan length of 2mm at nozzle-substrate separation of 70mm at 20°C. The
29
30
31 profilometer scan with S= 50mm/s (Figure 8c) is merely a line over few discontinuous droplets of
32
33 sprayed polymer rather than a film as it is formed in case of S= 10mm/s and S= 30mm/s (Figure 8a and
34
35 8b respectively). Figure 4 and Figure 8 show that the thickness of the film formed at 20°C with S=
36
37
38 30mm/s and 50mm/s is below our aim of 500nm-1µm. In order to achieve a continuous film with
39
40 uniform thickness in the mentioned range and with a low roughness a second set of experiments was
41
42
43
performed with a speed of spray 10mm/s at 20°C changing the nozzle substrate distance in small steps
44
45 from N= 75mm to N= 45mm.
46
47 Figure 9 presents the microscope images and respective profilometer scans at variable nozzle-substrate
48
49
50 distances. It is seen that when the distance between nozzle and substrate is high (75mm and 70mm,
51
52 Figure 9a and 9b respectively), the surface of the film is dotted with dried droplets which is reflected in
53
54 a quite high Ra value of 180nm. When moving close to the substrate the opposite phenomenon is
55
56
57
58 18
59
60
1
2
3
4
5 observed. The “coffee ring effect” starts to appear when N= 60mm (Ra= 69nm in between the peaks,
6
7 Figure 9d). As the nozzle moves even closer to the substrate (N= 45mm, Figure 9e) the effect becomes
8
9 more prominent, edge peaks appear at the edge of the coated area and the film is very uniform in
10
11
12 between with Ra= 22nm. At high N the sprayed droplets are drying on the way to reach the substrate.
13
14 Here spray coating is performed in the dry regime. The same effect is observed when the temperature
15
16 of the substrate is 60°C (Figure 7d) while spraying and the droplets are drying instantaneously after
17
18
19 they reach the substrate. At low N the polymer is drying slowly which allows it to have enough
20
21 mobility to form edge peaks at the side of the coated area.
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52 Figure 9. OM image and profilometer scan showing the effect of distance between nozzle and substrate
53 (a) 75mm , (b) 70mm, (c) 65mm, (d) 60mm and (e) 45mm at 10mm/s and 20°C
54
55
56
57
58 19
59
60
1
2
3
4
5 The optimized state at N= 65mm (Figure 9c) lies in the narrow window between these two regimes of
6
7 dry and wet where we find a compromise between no edge peaks and a low Ra (76nm).
8
9
10 Influence of substrates
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38 Figure 10. OM image of PVP spray coated from aqueous solution on (a) Glass, (b) Silicon dioxide, (c)
39
40 SU-8, (d) HF etched silicon (e) Fluorocarbon coated Silicon and (f) spray coated from dichloromethane
41 solution on glass at T=20°C, S=10mm/s and N=45mm
42
43 Although we have used microscope glass slides as substrate for the study it is interesting to see the
44
45 effect of substrates on spray coating. The surface energy of the substrate plays an important role in
46
47
48
determining the quality of the film formed on it8. Different substrate materials frequently used in
49
50 microfabrication spanning a large range of surface energies were selected. It is seen that uniform films
51
52 are formed on glass and silicon (Figure 10a and b) whereas the films formed on SU-8, HF etched
53
54
55 silicon (hydrophobic silicon) and fluorocarbon coated silicon (Figure 10c, d and e) are discontinuous.
56
57
58 20
59
60
1
2
3
4
5 This is explained by the higher surface free energy of glass [water contact angle (θ) ~27°25] and silicon
6
7 dioxide (θ~45°26) compared to SU-8 (θ~78°27), HF etched silicon (hydrophobic silicon) (θ~70°26) and
8
9
10 fluorocarbon coated silicon (θ~106°27) which have low surface energy preventing adequate wetting of
11
12 the surface.
13
14
15 Influence of solvents
16
17
18
19 To verify our understanding of the influence of rapid drying of droplets on surface morphology, a
20
21 solution of PVP in dichloromethane was prepared and compared to the one obtained with PVP with
22
23 water as solvent. The parameters for the test were T= 20°C, N= 45mm and S= 10mm/s. With both
24
25
26 solvents a continuous film is formed, suggesting negligible differences in surface wetting. From Figure
27
28 10f, it is concluded that dichloromethane is showing the typical dry regime behavior with Ra =560nm
29
30 and no edge peaks. At the same condition PVP with water as solvent is forming a uniform film (Ra=
31
32
33 22nm) but with edge peaks (Figure 10a). Dichloromethane has a low boiling point of 40°C and hence a
34
35 higher evaporation rate than water resulting in a deposition in the dry regime. This demonstrates that
36
37
38
additionally to temperature and nozzle-substrate distance, the solvent selected to prepare the solution
39
40 can also tune the dry and wet regime deposition. For inkjet printing, extensive work has been reported
41
42 in minimizing the edge peak formation and extension of the narrow regime of uniform film formation
43
44
45 by using co-solvent systems28–30. The experiments with different solvents indicate that this also might
46
47 be applicable for spray coating. However, in this work we focused mainly on optimization of the
48
49 processing parameters of spray coating to find the regime of uniform film formation.
50
51
52
53 Spray Coating with PDLLA
54
55
56
57
58 21
59
60
1
2
3
4
5 Extensive studies with the PVP-water system have been used to build up knowledge on the role of
6
7 different spray coating parameters and their interrelations. A less detailed study with PDLLA in
8
9 cyclohexanone is performed and the observations are compared with the ones obtained during the
10
11
12 experiments with PVP.
13
14
15 Influence of solvents
16
17
18 In the case of the PVP, solvent selection plays an important role in final quality of the film. Therefore
19
20
solutions of PDLLA in cyclohexanone, toluene and dichloromethane were prepared and their film
21
22
23 forming behavior at room temperature was compared. It is seen from the OM images in Figure11 that
24
25 at identical spraying conditions the film formed with cyclohexanone (Figure11a) is more uniform than
26
27
28 the one formed with toluene (Figure 11b) and dichloromethane (Figure 11c). The observed differences
29
30 in film morphology could be due to the different evaporation rates or different surface tensions of the
31
32 three solvents. All the solutions showed good wetting of the surface forming a continuous film.
33
34
35 Therefore it is assumed that differences in the evaporation rate are the predominant effect.
36
37 Dichloromethane and toluene evaporated too fast resulting in a rough surface similar to the one
38
39 observed for PVP deposition with dichloromethane (Figure 10f). Hence cyclohexanone was selected as
40
41
42 the solvent for PDLLA for the study.
43
44
45
46
47
48
49
50
51
52
53 Figure 11. OM images of film formed using (a) cyclohexanone, (b) toluene and (c) dichloromethane as
54
55
solvent for spray coating PDLLA at 20°C with speed 10mm/s and nozzle-substrate separation 45mm
56
57
58 22
59
60
1
2
3
4
5 Influence of distance between nozzle-substrate
6
7
8 Similar to spray coating of PVP, the distance between nozzle and substrate is the most significant
9
10 parameter influencing the deposition of PDLLA. Figure 12 presents the microscope images and
11
12
13
respective profilometer scans at variable nozzle-substrate distances. No true dry regime is observed but
14
15 a uniform film formation does not occur (fig.12a, b) unless an optimum N=45mm is reached (fig.12c).
16
17 Due to the high boiling point of the solvent cyclohexanone, the PDLLA has sufficient mobility after
18
19
20 spraying to form puddles of polymer at the edge of the coated areas. This results in the formation of
21
22 edge peaks. For PDLLA, formation of edge peaks could not be completely avoided. Nevertheless, a
23
24 satisfying result is achieved as the peaks are narrow in width and the height difference between the
25
26
27 center of deposition and the peak is small. Furthermore, the roughness in Figure 12c is very low (Ra=
28
29 21 nm) for a thickness of 430 nm.
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58 23
59
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29 Figure 12. OM image and profilometer scan at nozzle-substrate distance of (a) 75mm, (b) 60mm and
30
(c) 45mm at 20°C and speed of spray 10mm/s
31
32
33 Influence of temperature and speed of the spraying nozzle
34
35
36 The influence of temperature and speed of the spraying nozzle on the surface morphology of the
37
38 sprayed film is quite similar to the PVP-water system. Figure 13 shows the microscope images and
39
40
41 respective profilometer scans at variable temperature of the substrate and speed of spray. The film
42
43 formed at 20°C with a speed of spray 10mm/s (Figure13a) is uniform compared to the film formed at
44
45 elevated temperatures (Figure13b, c) and at a higher speed of 50mm/s (Figure13d). Similar to the
46
47
48 nozzle-substrate distance, no true dry regime is found but due to an increase in the rate of drying at
49
50 higher temperature non-uniformity in the film is clearly observed both by visual inspection in OM
51
52
images and their corresponding profilometer scans. (Figure 13b, c). At high speed of the spraying
53
54
55 nozzle the amount of material deposited is not adequate to form a uniform film. The discontinuity of
56
57
58 24
59
60
1
2
3
4
5 the film can be observed from the OM image and profilometer scan in Figure 13d.The thickness of the
6
7 film increases when the speed of spray is decreased from 50mm/s (Figure 13d) to 10mm/s (Figure13a).
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37 Figure 13.OM image and profilometer scan at (a) 20°C, (b) 40°C, (c) 80°C when nozzle-substrate
38 distance is 45mm and speed of spray 10mm/s and (d) at speed of spray 50mm/s when nozzle-substrate
39
40 distance is 45mm at 20°C.
41
42
43
44 CONCLUSION
45
46 This study demonstrates the influence of different factors on homogeneous film formation by spray
47
48
49 coating. Here, we found the strength of full factorial experimental design and subsequent analysis in
50
51 determining the critical parameters of spray operation to achieve a continuous film with uniform
52
53
thickness and low roughness. Depending on the drying of the sprayed drops on the substrate we define
54
55
56 two broad regimes, ‘dry’ and ‘wet’. In the wet regime formation of edge peaks is observed which
57
58 25
59
60
1
2
3
4
5 should be avoided. In the dry regime, no edge peaks are formed, but deposition results in a rough
6
7 surface. The optimum condition of spraying lies in a narrow window between these two regimes where
8
9 we obtain a film of low roughness and with no edge peaks. Thus to move from a wet to a dry regime
10
11
12 three possibilities can be summarized as 1) an increase in temperature to accelerate drying of the
13
14 droplets deposited on the substrate, 2) the selection of a solvent with higher evaporation rate for
15
16 spraying, 3) an increase of the distance between nozzle and substrate where drying of the droplets
17
18
19 before reaching the substrate is enhanced. For both PVP and PDLLA spray coating, the distance
20
21 between nozzle and substrate is the most significant parameter. Both low speed of the spraying nozzle
22
23
24
and substrate with high surface energy result in formation of a continuous film. Even though we are
25
26 working with two completely different systems, the experiments with PDLLA confirm the observations
27
28 found for PVP. For both cases optimum spraying temperature is 20°C and low speed of spray of
29
30
31 10mm/s. The nozzle-substrate distance is fine tuned for the respective systems and the optimum
32
33 distance is 65mm for PVP and 45mm for PDLLA. Continuous film formation occurs on high surface
34
35 energy glass surface with the high boiling point solvents water for PVP and cyclohexanone for
36
37
38 PDLLA. The detailed study of the different process parameters with two different polymer-solvent
39
40 systems improves the understanding of the technique of spray coating. The results can be generalized
41
42 and other systems can benefit from the conclusions of this work in the future.
43
44
45
46 AUTHOR INFORMATION
47
48
49 Corresponding Author
50
51
52 *Email: Stephan.Keller@nanotech.dtu.dk
53
54
55 Notes
56
57
58 26
59
60
1
2
3
4
5 The authors declare no competing financial interest
6
7
8 ACKNOWLEGDEMENTS
9
10
11 This research is supported by the Villum Kann Rasmussen Centre of Excellence “NAMEC” under
12
13 Contract No.65286
14
15
16
17 REFERENCE
18
19 1. Krebs, F. C. Fabrication and Processing of Polymer Solar Cells: A Review of Printing and Coating
20 Techniques. Sol. Energ.Mat.Sol. C 2009, 93, 394–412.
21
22
2. Lu, W.; Fadeev, A. G.; Qi, B.; Smela, E.; Mattes, B. R.; Ding, J.; Spinks, G. M.; Mazurkiewicz, J.; Zhou,
23
24
D.; Wallace, G. G. et al. Use of Ionic Liquids for Pi-conjugated Polymer Electrochemical Devices.
25 Science 2002, 297, 983–987.
26
27 3. Panasyuk, T. L.; Mirsky, V. M.; Piletsky, S. A.; Wolfbeis, O. S. Electropolymerized Molecularly
28 Imprinted Polymers as Receptor Layers in Capacitive Chemical Sensors. Anal. Chem. 1999, 71, 4609–
29 4613.
30
31 4. Lochon, F.; Fadel, L.; Dufour, I.; Rebière, D.; Pistré, J. Silicon Made Resonant Microcantilever:
32
Dependence of the Chemical Sensing Performances on the Sensitive Coating Thickness. Mat. Sci. Eng. C
33
34
2006, 26, 348–353.
35
36 5. Vendra, V. K.; Wu, L.; Krishnan, S. Polymer Thin Films for Biomedical Applications. In Nanostructured
37 Thin Films and Surfaces; Kumar, C. S. S. R., Ed.; Wiley-VCH: Weinheim, 2010; Vol. 5, pp. 1–51.
38
39 6. Wang, M.; Yang, G.-Z.; Wang, M.; Liu, T. Effect of Film Thickness Controlled by Ink-jet Printing
40 Method on the Optical Properties of an Electroluminescent Polymer. Polym. Advan. Technol. 2010, 21,
41 381–385.
42
43
44 7. Keller, S. S.; Gammelgaard, L.; Jensen, M. P.; Schmid, S.; Davis, Z. J.; Boisen, A. Deposition of
45 Biopolymer Films on Micromechanical Sensors. Microelectron. Eng. 2011, 88, 2297–2299.
46
47 8. McGrath, M. G.; Vrdoljak, A.; O’Mahony, C.; Oliveira, J. C.; Moore, A. C.; Crean, A. M. Determination
48 of Parameters for Successful Spray Coating of Silicon Microneedle Arrays. Int. J. Pharm 2011, 415,
49 140–149.
50
51
9. Linden, J.; Thanner, C.; Schaaf, B.; Wolff, S.; Lägel, B.; Oesterschulze, E. Spray Coating of PMMA for
52
53
Pattern Transfer via Electron Beam Lithography on Surfaces with High Topography. Microelectron. Eng.
54 2011, 88, 2030–2032.
55
56
57
58 27
59
60
1
2
3
4
10. Yu, L.; Lee, Y. Y.; Tay, F. E. H.; Iliescu, C. Spray Coating of Photoresist for 3D Microstructures with
5
6 Different Geometries. JPCS 2006, 34, 937–942.
7
8 11. Pham, N. P.; Burghartz, J. N.; Sarro, P. M. Spray Coating of Photoresist for Pattern Transfer on High
9 Topography Surfaces. J. Micromech. Microeng. 2005, 15, 691–697.
10
11 12. Pan, C. J.; Tang, J. J.; Weng, Y. J.; Wang, J.; Huang, N. Preparation, Characterization and
12 Anticoagulation of Curcumin-eluting Controlled Biodegradable Coating Stents. J. Control. Release 2006,
13 116, 42–49.
14
15
16 13. Tobiska, S.; Kleinebudde, P. Coating Uniformity: Influence of Atomizing Air Pressure. Pharm. Dev.
17 Technol. 2003, 8, 39–46.
18
19 14. Lang, R. J. Ultrasonic Atomization of Liquids. J. Acoust. Soc. of Am. 1962, 34, 6–8.
20
21 15. Pham, N. P.; Boellaard, E.; Burghartz, J. N.; Sarro, P. M. Photoresist Coating Methods for the Integration
22
of Novel 3-D RF Microstructures. J. Microelectromech. S. 2004, 13, 491–499.
23
24
25 16. Perfetti, G.; Alphazan, T.; Hee, P. van; Wildeboer, W. J.; Meesters, G. M. H. Relation Between Surface
26 Roughness of Free Films and Process Parameters in Spray Coating. Euro. J. Pharm Sci. 2011, 42, 262–
27 272.
28
29 17. Deegan, R.; Bakajin, O.; Dupont, T. Capillary Flow as the Cause of Ring Stains from Dried Liquid
30 Drops. Nature 1997, 389, 827–829.
31
32
33
18. Dubourg, G.; Fadel-Taris, L.; Dufour, I.; Pellet, C.; Ayela, C. Collective Fabrication of All-organic
34 Microcantilever Chips Based on a Hierarchical Combination of Shadow-masking and Wafer-bonding
35 Processing Methods. J. Micromech. Microeng. 2011, 21, 095021.
36
37 19. Gadelmawla, E. S.; Koura, M. M.; Maksoud, T. M. a.; Elewa, I. M.; Soliman, H. H. Roughness
38 Parameters. J. Mater. Process. Tech. 2002, 123, 133–145.
39
40 20. http://people.revoledu.com/kardi/tutorial/SVM/What-is-SVM-An-intuitive-introduction.html.
41
42
43 21. Http://www.mathworks.se/help/stats/stepwisefit.html.
44
45 22. Fleury, B.; Dantelle, G.; Darbe, S.; Boilot, J. P.; Gacoin, T. Transparent Coatings Made from Spray
46 Deposited Colloidal Suspensions. Langmuir 2012, 28, 7639–7645.
47
48 23. Kang, H.; Soltman, D.; Subramanian, V. Hydrostatic Optimization of Inkjet-printed Films. Langmuir
49 2010, 26, 11568–11573.
50
51
52 24. Wrobel, G. Optimization of High-density cDNA-microarray Protocols by “Design of Experiments”.
53 Nucleic Acids Res. 2003, 31, e67.
54
55
56
57
58 28
59
60
1
2
3
4
25. Canning, J.; Petermann, I.; Cook, K. Surface treatment of silicate based glass: base Piranha treatment
5
6 versus 193nm laser processing. In Proc. of SPIE; Canning, J.; Peng, G., Eds.; 2012; Vol. 8351, p.
7 83512N–83512N–5.
8
9 26. Franssila, S. Introduction to Microfabrication; Second.; John Wiley & Sons, Ltd: West Sussex, 2010; p.
10 150.
11
12 27. Keller, S. Fabrication of an Autonomous Surface Stress Sensor with the Polymer SU-8, Technical
13 University of Denmark, 2008.
14
15
16 28. Oh, Y.; Kim, J.; Yoon, Y. J.; Kim, H.; Yoon, H. G.; Lee, S.-N.; Kim, J. Inkjet Printing of Al2O3 Dots,
17 Lines, and Films: From Uniform Dots to Uniform Films. Curr. Appl . Phys. 2011, 11, S359–S363.
18
19 29. Teichler, A.; Eckardt, R.; Hoeppener, S.; Friebe, C.; Perelaer, J.; Senes, A.; Morana, M.; Brabec, C. J.;
20 Schubert, U. S. Combinatorial Screening of Polymer:Fullerene Blends for Organic Solar Cells by Inkjet
21 Printing. Adv. Energy Mater. 2011, 1, 105–114.
22
23
30. Teichler, A.; Perelaer, J.; Schubert, U. S. Screening of Film-Formation Qualities of Various Solvent
24
25 Systems for π-Conjugated Polymers Via Combinatorial Inkjet Printing. Macromol. Chem. Physic. 2013,
26 214, 547–555.
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58 29
59
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16 Table of contents only
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58 30
59
60

Process Optimization of Ultrasonic Spray Coating of Polymer Films

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Process Optimization of Ultrasonic Spray Coating of Polymer Films

Uploaded by

Copyright:

Available Formats

Subscriber access provided by HARVARD UNIVERSITY

You might also like