biosensors Article
Improved Manufacturing Performance of Screen Printed Carbon Electrodes through Material Formulation Eifion Jewell *, Bruce Philip and Peter Greenwood SPECIFIC, College of Engineering, Swansea University, Fabian Way, SA1 8AN Swansea, UK;
[email protected] (B.P.);
[email protected] (P.G.) * Correspondence:
[email protected]; Tel.: +44-179-229-6395 Academic Editors: Donatella Albanese and Roberto Pilloton Received: 21 April 2016; Accepted: 16 June 2016; Published: 27 June 2016
Abstract: Printed carbon graphite materials are the primary common component in the majority of screen printed sensors. Screen printing allows a scalable manufacturing solution, accelerating the means by which novel sensing materials can make the transition from laboratory material to commercial product. A common bottleneck in any thick film printing process is the controlled drying of the carbon paste material. A study has been undertaken which examines the interaction between material solvent, printed film conductivity and process consistency. The study illustrates that it is possible to reduce the solvent boiling point to significantly increase process productivity while maintaining process consistency. The lower boiling point solvent also has a beneficial effect on the conductivity of the film, reducing the sheet resistance. It is proposed that this is a result of greater film stressing increasing charge percolation through greater inter particle contact. Simulations of material performance and drying illustrate that a multi layered printing provides a more time efficient manufacturing method. The findings have implications for the volume manufacturing of the carbon sensor electrodes but also have implications for other applications where conductive carbon is used, such as electrical circuits and photovoltaic devices. Keywords: screen printing; carbon electrodes; manufacturing
1. Introduction Screen printed carbon graphite pastes are the main materials used for bio sensors electrodes, [1–4]. From a research perspective carbon/graphite materials are considered commercially mature. In recent times little scientific research has been conducted on the impact of formulation on sensor performance, although comparisons between commercial materials have been carried out [5,6] which highlighted each material behaved differently. More recently, research focus has concentrated on more research attractive materials such as Carbon nano tubes CNTs [7], and multi walled carbon nano tubes MWCNTs, Graphene and their hybridization with conventional graphite materials, [8]. The limited formulation studies have focused on the influence of the carbon/graphite relative proportions, the carbon/graphite physical and electrical characteristics and the proportion of conductive material required for charge percolation through the cured film [9]. A variety of binder systems are compatible with carbon/graphite materials, such as, epoxies, alkyds, acrylics, vinyls, polyurethane resins or phenolic resins and the choice is governed by the curing temperature, substrate which is being printed, film flexibility and the resin compatibility with the final application. The choice of binder/binders is usually proprietary information held by manufacturers. The beauty of the screen printing process is that it is equally suited to small batch laboratory testing as to volume manufacturing, making the process of upscaling manufacture less problematic.
Biosensors 2016, 6, 30; doi:10.3390/bios6030030
www.mdpi.com/journal/biosensors
Biosensors 2016, 6, 30
2 of 10
However, there remains significant challenges in transitioning a laboratory to manufacturing process and the solvent choice plays an important role in this transition. The role of the solvent is to dissolve the binder, control the formulation viscosity and control the drying and curing process. A common bottleneck in many printing processes is the drying and curing process, which has a subsequent effect on productivity and ultimately profitability. Given the thick film nature of the screen printing process, a more volatile solvent would seem preferable as it potentially improves process productivity. A more volatile solvent may however also result in reduced process consistency as it evaporates during the printing processing. This can lead to “drying in”, the process where the ink begins to dry in the mesh, which is a particular issue for finer detail printing. A more volatile solvent also leads to increased viscosity variation and as the process is viscosity dependent [10], this can lead to variation in print quality. A further solvent effect on process is that the migration of the solvent into the polyurethane squeegee can impact the squeegee properties and hence impact process consistency [11,12]. During material formulation, a compromise must be made between solvent volatility, which is essentially a balance between productivity (which demands low boiling points) versus consistency (which requires a high boiling point). There is therefore an interaction between solvent properties, printed electrode characteristics and process productivity at volume. The solvent volatility has been shown to play role in film stresses induced during drying, [13], with faster evaporation leading to greater stress in the film [14]. The impact of this residual stress on conductivity of carbon materials has not been reported publically. In carbon graphite, there is a dearth of information on the effect of solvent on the performance of the cured film and the ease of processing. Given that the volume production is a primary advantage of screen printing, then this is worthy of investigation. The aim of the study was to understand the impact of solvent on the screen printed curing performance of the carbon graphite conductive nitrocellulose (NC) based materials and to establish the role of the solvent on performance, process consistency and stability. This information would allow material formulators to estimate the balance of solvents and allow sensor manufacturers to estimate the impact of formulation changes on their process. 2. Materials and Methods The materials were manufactured by Gwent Electronic Materials using the same carbon graphite/carbon black proportion, the same nitrocellulose binder and with the same solvent mass proportion. The net result was four materials whose constituents were identical in constituent parts apart from the solvent. The solvents and solvent blends were chosen from commercially viable solvents (health and safety compliant and cost) and on the basis of their relative boiling point. The boiling point being the characteristic that largely dictates the drying and in process variation. The final composition of the material is shown in Table 1. Table 1. Material used in the experimental study. I
II
III
IV
C: Binder: Solvent wt %
31:9:60
31:9:60
31:9:60
31:9:60
Solvents
4-Hydroxy-4-Methyl-2 Pentanone (Diacetone alcohol)
P-Menth-1-En-8-ol (Terpineol)
2-Butoxyethanol
4-Hyrdroxy-4Methyl-2Pentanone & P-Menth-1-En-8-ol (Terpineol)
Nominal solvent BP (˝ C)
166
219
171
166 & 219
Solvent viscosity (mPas) @20 ˝ C
2.9
36
2.9
2.9 & 36
Each material was printed through five screens (Table 2) in order to examine the interaction of the film thickness and solvent on the curing characteristics. The screens represent a range which is
Biosensors 2016, 6, 30
3 of 10
applicable to the production of carbon electrode sensors. The theoretical volume represents the geometric free air volume available within the mesh and is an indicator of film thickness only. The material was printed to a polyester coated steel substrate which was chosen as it provided a surface which was compatible with the material and general sensor manufacturing but also facilitated a structurally stable base substrate which would not undergo any form of deformation at high temperatures. Printing was carried out on an ATMA AT-25PA screen printer with drying taking place in a 3 m Thieme conveyor tunnel dryer operating at 1.2 m/min, a residence time of 2.5 min. An air temperature of 155 ˝ C was used to provide a favourable environment for substrate heating and mass transfer from the coating, with air impinged at 155 ˝ C to the surface from an array of holes. Multiple passes were made through the tunnel dryer with electrical performance being monitored at each intermediate stage on a cooled substrate. Table 2. Screens used in the experiment. Notation
32–100
61–64
77–48
90–48
110–34
Mesh ruling (threads/cm) Thread diameter (µm) Mesh aperture size(µm) Theoretical volume (cm3 /m2 )
32 100 209 73
61 64 90 30
77 48 77 27
90 48 55 19
110 34 54 19
Electrical performance was characterised by sheet resistance which was measured using 4-point sheet resistance with 2 mm electrode spacing and the subsequent sheet resistance was calculated using constants which reflect the geometric sizes being measured, [15]. Three measurements were made on each of five substrates. Within the sample size of 15 measurement points, the uncertainty in the sheet resistance was calculated to be ˘ 1 Ω/sq. In addition to the sheet resistance, two point resistance measurements were carried out on a spiral structure containing 56 cm of line of 700 µm width. In this way, the impact on fine structures, which are more relevant to sensor electrode design and manufacture, could also be analysed. Film thickness measurements were carried out using an Elcometer 456 induction gauge which was zeroed on the base polyester coating. Thermogravimetric analysis (TGA) was used to establish the relative evaporative characteristics of the materials once formulated, which when combined with rheological data would provide an indicator of manufacturing capability. TGA analysis was carried out using a Perkin Elmer Pyris 1 analysis in a nitrogen atmosphere. TGA isotherms were carried out at 130 ˝ C and 155 ˝ C as these represented the mean substrate temperature experienced during the substrate passage through the dryer (as measured by contact thermocouples) and the set air impingement temperature respectively. The true material exposure temperature therefore lies between these limits. Material rheology was carried out using a Brookfield RS cone and plate viscometer at 20 ˝ C operating through a constant stress shear ramp from 5 to 500 Pa with 50 intermediate steps. 3. Results 3.1. Material Characteristics The impact of the solvent on the nominal drying characteristics (as measured by TGA) is shown in Figure 1. Where there is a constituent of lower boiling point material, (I, II and IV), all solvent is lost (40% solids) by approximately 15 min at 130 ˝ C and by 6 min at 155 ˝ C. The higher boiling point material (II) does not achieve its stable solids level of 40% within 30 min at 120 ˝ C and takes approximately 10 min to achieve the solids level at 155 ˝ C. The solvent has a minimal effect on the form of the conductive paste viscosity profile, Figure 2. Each exhibits pseudoplastic behaviour with low shear viscosities upwards of 40 PaS. Given that there is an order of magnitude difference in the viscosity of the solvent (Table 2) and that this is the primary constituent of each material (60%), then the viscosity range of 5 PaS can be considered relatively small
Biosensors 2016, 6, 30
4 of 10
100% 100% 100% 90% 90% 90%
80% 80% 80% 70% 70% 70%
80% 80% 70% 70%
60% 60% 60% 50% 50% 50%
II I
40% 40% 40% 30% 30% 30%
IIIIII III
% % Initial Initial mass mass Initial
100% 100% 100% 90% 90% 90%
% Initial mass
% % Initial Initial mass mass
in comparison to the reduction in viscosity with increasing shear. Within this range the materials containing the high viscosity Terpineol solvent (II & IV) produce the highest viscosities while the lowest viscosity is306,obtained with lowest viscosity solvent (III). Biosensors 2016, 6, of 10 Biosensors 2016, 30 4 of4 10
IIIIII IVIV IV
20% 20% 20% 10% 10% 10% 0% 0% 0% 0 00
55
5
10 10 15 15 15 20 20 20 10 Time (min) Time (min) (min) Time
2525 25
60% 60% 50% 50%
I
II
40% 40% 30% 30%
II
IIII
III III III IV IV IV
20% 20% 10% 10% 10% 0% 0% 0%
3030 30
00
55
0
10 10
5
10
(a) (a)
15
15 15 Time (min) (min) Time
Time (min)
2020
20
25 25
25
30 30 30
(b) (b)
(b)
Figure 1. Thermogravimetric analysis (TGA) analysis of the materials at an isothermal temperature Figure 1. Thermogravimetric analysis(TGA) (TGA)analysis analysis of of the the materials materials atatan temperature Figure 1. Thermogravimetric analysis anisothermal isothermal temperature of of (a) Mean substrate temperature of 130 °C and (b) Impingement air temperature of 155 °C.
of (a)substrate Mean substrate temperature of 130 °C and Impingementair air temperature temperature ofof155 ˝ C and ˝ C. (a) Mean temperature of 130 (b)(b) Impingement 155°C. 40
40
35
30 25
Viscosity (Pas)
Viscosity Viscosity (Pas) (Pas)
35 30 I
25 20
II I
20 15
IIIII
15 10
IVIII
10 5
IV
5 0 0
0 0
100
100
200
200
400
500
600
300 400 -1 Shear Rate (s-1 )
300
500
600
Shear Rate (s-1 )
Figure 2. Controlled stress viscosity profiles for each material (50 points on each curve).
Figure 2. Controlled stressviscosity viscosity profiles profiles for (50(50 points on each curve). Figure 2. Controlled stress foreach eachmaterial material points on each curve).
One of the primary impacts of a higher volatility solvent is to accelerate the rate of viscosity variation due to solvent evaporation. In order to simulate the effect of changes in material rheology One of the primary impactsofofaahigher higher volatility volatility solvent to to accelerate the the rate rate of viscosity One of the impacts solventiswere is accelerate viscosity during theprimary printing process, known quantities of each material left to evaporate solventof and variation due to solvent evaporation. In order to simulate the effect of changes in material rheology variation due to in solvent evaporation. simulate the effect of changes in material rheology variations mass and viscosity dueIn toorder solventtoloss were measured at regular intervals. As the form during the printing process, known quantities of each material were left to evaporate solvent and viscosity/shear rate curve isquantities similar the of impact solvent loss is best by studying during of theeach printing process, known eachofmaterial were leftexamined to evaporate solvent and variations in massaand viscosity due to100 solvent lossmaterial were measured at regular intervals. Asbetween the form the in viscosity shear rate. s−1, each shows a near linear relationship variations mass at andgiven viscosity dueisAt to solvent loss were measured at regular intervals. As the form of each viscosity/shear rate curve similar the impact of solvent loss is best examined by studying the solvent lost and the viscosity, Figure 3. The limit of 3% mass loss was imposed as the samples of −1 the impact of solvent loss is best examined by studying of each viscosity/shear rate curve is similar the higher viscosity at a given shear rate. 100tos−1 , eachincreased material shows between viscosity material were At felt have beyonda near that linear whichrelationship could be printed ´1 , each material shows a near linear relationship between the viscosity at lost a given shear rate. At 100 s thesuccessfully. solvent and the viscosity, Figure 3. The limit of 3% mass loss was imposed as the samples Although the behaviour of each material is similar, clearly the time required to achieve of higher viscosity material were felt to increased that which could be the solvent lost and 3. have The limit of 3%beyond mass loss was imposed theprinted samples of the same massthe lossviscosity, would beFigure lower for the more volatile materials, for a given set ofasoperating successfully. Although the behaviour of each material is similar, clearly the time required to achieve conditions.material were felt to have increased beyond that which could be printed higher viscosity successfully.
the same mass loss would lower for moreclearly volatilethe materials, for a given set of the operating Although the behaviour of eachbematerial is the similar, time required to achieve same mass conditions. 25 loss would be lower for the more volatile materials, for a given set of operating conditions. 24
25 Viscosity (PaS)
24
Viscosity Viscosity (PaS) (PaS)
23 22
23 22 21
I
20
II
19
21 18
III
20 17
IV
I II
19 16
III
18 15 0
17
0.5
1
1.5
2
2.5
2
2.5
3
IV
Solvent loss %
16 15 0
0.5
1
1.5
3
0 0.5 1 1.5 2 2.5 3 Figure 3. Simulation of the effect of solvent loss on material viscosity at a shear rate of 100 s−1. Solvent loss %
Figure 3. Simulation of the effect of solvent loss on material viscosity at a shear rate of 100 s´1 . −1 −1 Figure 3. Simulation of the effect of solvent loss on material viscosity at a shear rate of 100 s .
Biosensors 2016, 6, 30
5 of 10
Biosensors 2016, 6, 30
5 of 10
3.2. Printed Characteristics
3.2. Printed Characteristics Biosensors 2016, 6, 30 of 10 Having established the properties of the materials, each sample was subjected to an5identical Having established the properties of theexhibiting materials, each sample was subjected to an identical printing process. All materials printed well, excellent adhesion and dried without any 3.2. Printed Characteristics printing process. All materials printed well, exhibiting excellent adhesion and dried without any used defects such film cracking, delamination or mottling. The interaction between the mesh type Having established the properties of the materials, each sample was subjected to an identical defects such film cracking, delamination or mottling. The interaction between the mesh type used and the solvent volatility on the curing time and electrical performance of the cured film is shown in printing process. All materials printed well,and exhibiting excellent adhesion and dried without any and solvent volatility on the mesh curing110–34 time electricalthe performance of the cured filmalthough is shown in each Figure 4. the In all instances the finest produces highest sheet resistance, in defects 4. such film cracking, or mottling. The interaction the meshalthough type used Figure In all instances thedelamination finest mesh 110–34 produces the highest between sheet resistance, in instance the change in sheet resistance is minimal with increased residence time (successive passes and the solventthe volatility curing time and electricalwith performance the cured time film is shown in each instance changeon in the sheet resistance is minimal increasedofresidence (successive through the4.dryer). As the meshfinest coarseness increases the sheethighest resistance drops and the time taken Figure In all instances meshcoarseness 110–34 produces sheet resistance, although in to passes through the dryer).the As the mesh increasesthe the sheet resistance drops and the time achieve the stable sheet resistance increases. In all cases, the coarsest mesh (32–100) does not achieve a each instance thethe change sheet resistance is minimal increased residence time (successive taken to achieve stable in sheet resistance increases. In allwith cases, the coarsest mesh (32–100) does not stableachieve valuethrough a residence 530 time s.coarseness passes the dryer). As theofmesh awith stable value withtime a residence of 530 s.increases the sheet resistance drops and the time
35 50 30 45
32-100 61-64
25 40 20 35
77-48 32-100 90-48
15 30 10 25
61-64 110-34 77-48
5 20 0 15 0
10
100
200
300
0 0
100
200
500
600
(a) 400
300
45
110-34
500
77-48 32-100 90-48
15 30 10 25
61-64 110-34 77-48
5 20 0 15 300
400
500
600
0
100
200
61-64 110-34 77-48
5 20 0 15 10
0
100
200
300
(c) 400
300
90-48
100
200
500
90-48
600
110-34
(b)400
300
500
600
Residence time (s)
(b)
40 35 50 30 45
32-100 61-64
25 40 20 35
77-48 32-100 90-48
15 30 10 25
61-64 110-34 77-48
5 20 0 15 10
0
100
200
300
600
400
500
90-48
600
110-34
Residence time (s)
5 0
500
400
Residence time (s)
5
110-34
Residence time (s)
0
77-48 32-100 90-48
15 30 10 25
0
61-64
200
61-64
45
25 40 20 35
100
32-100
25 40 20 35
50
32-100
0
35 50 30 45
600
35 50 30 45
5
40
0
(a)
40
10
90-48
Residence time (s)
50
Sheet (Ω /sq) Sheet Resistane (ΩResistane /sq)
400
Residence time (s)
5
Sheet (Ω /sq) Sheet Resistane (ΩResistane /sq)
40
Sheet (Ω /sq) Sheet Resistane (ΩResistane /sq)
Sheet (Ω /sq) Sheet Resistane (ΩResistane /sq)
taken to achieve the stable sheet resistance increases. In all cases, the coarsest mesh (32–100) does not achieve a 50 stable value with a residence time of 530 s. 50 45 45
0
100
Residence time (s)
200
(d)400
300
500
600
Residence time (s)
Figure 4. Sheet resistance/residence time profiles for each solvent (I to IV) printed through each
Figure 4. Sheet resistance/residence time profiles for each solvent (I to IV)(d) printed through each screen. (c) screen. (a) I; (b) II; (c) III; (d) IV. (a) I; (b) II; (c) III; (d) IV. Figure 4. Sheet resistance/residence time profiles for each solvent (I to IV) printed through each
12 14 10
32-100
12 8
61-64
10 6
77-48 32-100 90-48 61-64 110-34 77-48
8 4 6 2
90-48
4 0 2
110-34 I
II
III
IV
(a)
0 I
II
Resultant Sheet Resistance (Ω /aq) Resultant Sheet Resistance (Ω /aq)
Change in Sheet Resistance (Ω /aq) Change in Sheet Resistance (Ω /aq)
The interaction the solvent type and the change in resistance over the maximum screen. (a) I; (b) II; between (c) III; (d) IV. residence time of 530 s is summarised Figure final in resistance achieved after the maximum The interaction between the solvent in type and 5a. theThe change resistance over the maximum residence The interaction between the solvent type and the change in resistance over the maximum residence time illustrates that the coarsest (32–100) produces the lowest sheet resistance and that, in time of 530 s is summarised in Figure 5a. The final resistance achieved after the maximum residence residence time of 530 s is summarised in Figure 5a. The final resistance achieved after the maximum instances, resistance increases in a step manner the mesh becomes time most illustrates thatthe thesheet coarsest (32–100) produces the wise lowest sheetasresistance and that,finer, in most residence illustrates that the coarsestis(32–100) produces sheet resistance Figure 5b.time The change in sheet resistance most evident withthe thelowest coarsest mesh (32–100).and that, in instances, the sheet resistance increases in a step wise manner as the mesh becomes finer, Figure 5b. most instances, the sheet resistance increases in a step wise manner as the mesh becomes finer, The change in sheet resistance is most evident with the 50coarsest mesh (32–100). 14 5b. The change in sheet resistance is most evident Figure with the coarsest mesh (32–100). 45 40 50 35 45 30 40 25 35 20 30 15 25 10 20 5 15 0 10
32-100 61-64 77-48 32-100 90-48 61-64 110-34 77-48 90-48 110-34 I
II
I
II
5
IV
IV
III
IV
(b)
0 III
III
Figure 5. (a) Change in sheet resistance and (b) Minimum sheet resistance for a residence time of 530 (a) (b) s for each material through each mesh. Figure 5. (a) Change in sheet resistance and (b) Minimum sheet resistance for a residence time of 530 Figure 5. (a) Change in sheet resistance and (b) Minimum sheet resistance for a residence time of 530 s s for each material through each mesh.
for each material through each mesh.
Biosensors 2016, 6, 30 Biosensors 2016, 2016, 6, 6, 30 30 Biosensors
6 of 10 of 10 10 66 of
Sheet Resistivity (Ω /sq/mil) Sheet Resistivity (Ω /sq/mil)
When the conductivity of the cured materials is normalised using standard industry notation When where the conductivity of μm), the cured materialstoisexamine normalised industry notation (Ω/sq/mil 1 mil = 25.4 it is possible the using impactstandard of the mesh on the sheet When the conductivity of the cured materials is normalised using standard industry notation (Ω/sq/mil where 16.mil = 25.4 μm),ofit the is possible to examine the and impact of theform mesh onconductive the sheet resistance, Figure If the deposit material was consistent the film and (Ω/sq/mil where 1 mil = 25.4 µm), it is possible to examine the impact of the mesh on the sheet resistance, Figure 6. Ifwere the depositthen of the was sheet consistent and the film be form and conductive element distribution thismaterial normalised resistivity should independent of the resistance, Figure 6. If the equal, deposit of the material was consistent and the film form and conductive element distribution were equal, then this normalised sheet resistivity should be independent ofsheet the mesh used. However, materials with the more volatile solvents (I, III and IV) exhibit lower element distribution were equal, then this normalised sheet resistivity should be independent of the mesh used.for However, materials the more solvents III and IV) exhibit sheet resistivity the coarsest screenswith (thicker films)volatile than with those (I, with It islower postulated mesh used. However, materials with the more volatile solvents (I, IIIfiner and screens. IV) exhibit lower sheet resistivity for the coarsest screens (thicker films) than with those with finer screens. It is postulated that this is associated with some filtering by the finer meshes which reduce the larger particle resistivity for the coarsest screens (thicker films) than with those with finer screens. It is postulated that this is associated with some filtering by the finer meshes which reduce the larger particle graphite thesome film. that this iscomponent associated in with filtering by the finer meshes which reduce the larger particle graphite graphite component in the film. component in the film. 14 14 12 12 10 10 8 8 6 6 4 4 2 2 0 0
32-100 32-100 61-64 61-64 77-48 77-48 90-48 90-48 110-34 110-34
I I
II II
III III
IV IV
Figure 6. Normalised sheet resistivity (expressed in Ω/sq/mil) for each mesh ink formulation. Figure 6. Normalised sheet resistivity (expressed in Ω/sq/mil) for each mesh ink formulation. Figure 6. Normalised sheet resistivity (expressed in Ω/sq/mil) for each mesh ink formulation.
30
30
30 25
30 25
25 20
32-100
20 15
32-100 61-64
15 10
61-64 77-48 77-48 90-48
10 5
90-48 110-34
5 0 0
110-34 0
100
200
300
400
500
600
0
100
Residence (s) 200 300 time 400
500
600
Resistane (Ω /cm) Resistane (Ω /cm)
Resistane (Ω /cm) Resistane (Ω /cm)
When the resistivity is subsequently compared to that of similar commercial materials, it is When isissubsequently to that of similar commercial materials, it is shown When the resistivity subsequently compared toover that of similar materials, it is shown thatthe theresistivity formulation presents an compared improvement that which iscommercial available commercially. A that the formulation presents an improvement over that which is available commercially. A review of shown that the formulation presents an improvement over that which is available commercially. A review of commercial materials indicates that the minimum available sheet resistivity is around 14 commercial materials that the minimum available sheet resistivity is reduction around 14 Ω/sq/mil. review of commercial materials indicates that the minimum available sheet is in around 14 Ω/sq/mil. The materialindicates described therefore represents approximately a 50%resistivity material The material described therefore represents approximately a 50% reduction in material resistivity. Ω/sq/mil. The material described therefore represents approximately a 50% reduction in material resistivity. While sheet resistance resistanceisisananexcellent excellent measure of material, mesh curing performance resistivity. While sheet measure of material, mesh and and curing performance most most electrodes are made geometric structures and thus study ofperformance carbon electrode While sheet resistance isdiscrete an discrete excellent measure of material, mesh and curingof most electrodes are made fromfrom geometric structures and thus a astudy carbon electrode performance needs examine the ofofsuch The electrodes arealso made fromto discrete geometric structures andstructures. thus a study of resistance/curing carbon electrode performance also needs to examine the printing printing such structures. The resistance/curing characteristics of the structures closely follows that which is seen with the sheet resistance, Figure 7. 7. performance also needs to examine the printing of such structures. The resistance/curing characteristics of the structures closely follows that which is seen with the sheet resistance, Figure Additional has effect resistance the finest mesh (110–34) but characteristics of thetime structures closely follows thatthe which is seenfor with sheet resistance, 7. Additional drying drying time has aa negligible negligible effect on on the resistance for thethe finest mesh (110–34)Figure but has has an increasing effect as the mesh coarseness (film thickness) is increased. The materials containing Additional drying time has a negligible effect on the resistance for the finest mesh (110–34) but has an increasing effect as the mesh coarseness (film thickness) is increased. The materials containing low boiling (I, III IV) the resistance, values typically an effectsolvents as the mesh coarseness (film thickness) is line increased. The with materials containing lowincreasing boiling point point solvents (I, III and and IV) produced produced the lowest lowest line resistance, with values typically between 10 Ω/cm and 20 Ω/cm for the coarsest and finest mesh. The higher boiling point material (II) low boiling point solvents (I, III and IV) produced the lowest line resistance, with values typically between 10 Ω/cm and 20 Ω/cm for the coarsest and finest mesh. The higher boiling point material (II) consistently reports the highest line resistance. between 10 Ω/cm and 20 Ω/cm for the coarsest and finest mesh. The higher boiling point material (II) consistently reports the highest line resistance. consistently reports the highest line resistance.
25 20
32-100
20 15
32-100 61-64
15 10
61-64 77-48 77-48 90-48
10 5
90-48 110-34
5 0 0
Residence time (s)
110-34 0
100
200
300
400
500
600
0
100
Residence (s) 200 300 time 400
500
600
Residence time (s)
(I) (I)
(II) (II)
Figure 7. Cont.
Biosensors Biosensors 2016, 2016, 6, 6, 30 30
77 of of 10 10 7 of 10
30
30
30 25
30 25
25 20
32-100
20 15
61-64 32-100 77-48 61-64 90-48 77-48 110-34 90-48
15 10 10 5 05
Resistane (Ω /cm) Resistane (Ω /cm)
Resistane (Ω /cm) Resistane (Ω /cm)
Biosensors 2016, 6, 30
100
0
100
200
300
400
Residence time (s) 200 300 400
500
32-100
20 15
61-64 32-100 77-48 61-64 90-48 77-48 110-34 90-48
15 10 10 5 05
110-34 0
0
25 20
600
110-34 0
100
0 500
600
0
100
Residence time (s)
200
300
400
500
Residence time (s) 200 300 400
500
600 600
Residence time (s)
(III) (IV) (III) (IV) Figure 7. Resistance (expressed in Ω/ linear cm) for a 700 μm line—curing time profile for each of the material to IV) and(expressed mesh combinations. Figure 7. 7.(IResistance Resistance in Ω/ Ω/ linear for each each of of the the Figure (expressed in linear cm) cm) for for aa 700 700 μm µm line—curing line—curing time time profile profile for material (I (I to to IV) IV) and and mesh mesh combinations. combinations. material
3.3. Process Consistency Characteristics 3.3. Process Consistency Characteristics As highlighted in the introduction, the primary effects of the solvent on the process consistency 3.3. Process Consistency Characteristics are associated with solvent evaporation the altering material and solvent into the As highlighted in the introduction, primary effectsrheology of the solvent on the absorption process consistency As highlighted in the introduction, the primary effects of the solvent on the process consistency squeegee. In order investigate the former, a short production of 50 samples ofabsorption each material are associated withtosolvent evaporation altering material rheology and solvent intowas the are associated with solvent evaporation altering material rheology and solvent absorption into the carried outInatorder a production rate the of 30 s (total processing time of 50 25 samples min). These were given was the squeegee. to investigate former, a short production of each material squeegee. In order to investigate the former, a short production of 50 samples of each material was maximum time of 530 through the dryer in a single retarding speed of the carried outresidence at a production rates of 30 s (total processing time pass of 25bymin). Thesethe were given the carried out at a production rate of 30 s (total processing time of 25 min). These were given the maximum dryer conveyor. The sheet resistance of every 5th sample measured the deviation from maximum residence time of 530 s through the dryer in awas single pass byand retarding the speed of the the residence time of 530 s through the dryer in a single pass by retarding the speed of the dryer conveyor. mean starting value is illustrated in Figure 8. There is no clear trend in the measured sheet dryer conveyor. The sheet resistance of every 5th sample was measured and the deviation from the The sheet resistance of every 5th sample was measured and the deviation from the mean starting value resistances withvalue the variation observed deviating no more thanclear ± 1.5trend Ω/sq in from nominal mean. mean starting is illustrated in Figure 8. There is no thea measured sheet is illustrated in Figure 8. There is no clear trend in the measured sheet resistances with the variation Given a measurement uncertainty of ± 1 deviating Ω/sq thenno this can than safely± be to be no moremean. than resistances with the variation observed more 1.5assumed Ω/sq from a nominal observed deviating no more than ˘ 1.5 Ω/sq from a nominal mean. Given a measurement uncertainty natural process fluctuation. Given a measurement uncertainty of ± 1 Ω/sq then this can safely be assumed to be no more than of ˘ 1 Ω/sq then this can safely be assumed to be no more than natural process fluctuation. natural process fluctuation. Deviation resistance (Ω /sq) Deviation from from initialinitial sheetsheet resistance (Ω /sq)
2 2 1.5 1.5 1 1 0.5
I
0.5 0 0 -0.5
III 0
10
20
30
40
50
III II
0
10
20
30
40
50
IV III
-0.5 -1
IV
-1 -1.5 -1.5 -2 -2
Sample number Sample number
Figure 8. Variation observed in a short pilot run of 50 sheets. Figure 8. Variation observed in a short pilot run of 50 sheets.
Figure 8. Variation observed in a short pilot run of 50 sheets.
The other of the on process consistency is associated with the absorption The otherprimary primaryimpact impact of solvent the solvent on process consistency is associated with the of the solvent into the polyurethane squeegee. This was established by cutting two fresh squeegee absorption of theprimary solvent into the polyurethane squeegee. This was established cutting two The other impact of the solvent on process consistency is by associated withfresh the lengths, the edge of the first was placed in 10 mm of wet conductive material for 2 h. At end squeegee lengths, edgeinto of the was placedsqueegee. in 10 mm This of wet conductive material for 2 two h.the Atfresh the absorption of the the solvent thefirst polyurethane was established by cutting of a of 2 ha period the squeegee which hadhad been placed in the conductive carbon material was used to end 2 lengths, h period theedge squeegee been placed in the conductive carbon material used squeegee the of thewhich first was placed in 10 mm of wet conductive material for 2was h. At the print five samples andand thenthen immediately replaced with the dry length, to printtoa further samples. to print samples immediately replaced the dry length, print a five further five end of a five 2 h period the squeegee which had been placedwith in the conductive carbon material was used Thus, any changes in deposit can be attributed directly to the changes in squeegee properties due to samples. Thus, any changes in deposit can be attributed directly to the changes in squeegee to print five samples and then immediately replaced with the dry length, to print a further five absorptiondue of materials in the ink. Exposure toink. the Exposure material resulted in a negligible difference over properties absorption materials the to the material a negligible samples. Thus,toany changesof in depositincan be attributed directly to theresulted changesin in squeegee the 2 h period, Figure 9. There is an increase in sheet resistance of around 3 Ω/sq. for the material difference 2 h period,ofFigure 9. There is ink. an increase in to sheet of around Ω/sq. for propertiesover due the to absorption materials in the Exposure the resistance material resulted in a3negligible formulated with the typewith II solvent. This is beyond that which can bewhich attributed error (˘ 1 Ω/sq) the material formulated the type II solvent. This is beyond that can be attributed errorand (± difference over the 2 h period, Figure 9. There is an increase in sheet resistance of around 3 Ω/sq. for 1the Ω/sq) and shows that there some the solvent and thecan polyurethane material. material formulated withisthe typeinteraction II solvent.between This is beyond that which be attributed error (± 1 Ω/sq) and shows that there is some interaction between the solvent and the polyurethane material.
Biosensors 2016, 6, 30
8 of 10
Biosensors 2016, 6, 30
8 of 10
shows that there is some interaction between the solvent and the polyurethane material. The exact The exact and chemical andmechanism physical mechanism ofchange this small not been investigated as the chemical physical of this small has change not beenhas investigated as the performance performance of this material was the poorest of the materials investigated (Figures 4–7) and as such of this material was the poorest of the materials investigated (Figures 4–7) and as such its use would its not be pursued. notuse be would pursued. 50 45
Sheet Resistance (Ω /sq)
40 35 30 25 20 15 10 5 0 I
II Fresh
III
IV
2hr exposure
Figure Figure 9. 9. Sheet Sheetresistance resistanceof ofeach eachmaterial materialfor foraafresh freshand andthen thenaasqueegee squeegeewhich whichhas hasbeen beenexposed exposedin in the material for 2 h. the material for 2 h.
4. 4. Discussion Discussion This studyhas hasraised raiseda number a number issues which are worthy of discussion. The This experimental experimental study of of issues which are worthy of discussion. The sheet sheet resistance is significantly lower than that achieved with comparable commercial materials resistance is significantly lower than that achieved with comparable commercial materials which have which have been at best 14 The Ω/sq/mil. The nitrocellulose to flex without been found to be found at bestto 14be Ω/sq/mil. nitrocellulose binder wasbinder able towas flexable without impact on impact on the substrate. the substrate. The The choice choice of of solvent solvent has has been been shown shown to to have have an an impact impact on on not not only only the the drying drying but but on on the the final final conductivity of the film. Examined from a sheet resistance viewpoint, it has been possible to lower conductivity of the film. Examined from a sheet resistance viewpoint, it has been possible to lower the the thick sheet resistance by 50% by changing from solvent to solvent III (Figure 5b) and the thick filmfilm sheet resistance by 50% by changing from solvent II toIIsolvent III (Figure 5b) and the thin thin resistance by 25% by changing solvent to solvent III (Figure is proposed film film sheetsheet resistance by 25% by changing fromfrom solvent II to IIsolvent III (Figure 5b). 5b). It is It proposed that that the mechanism foristhis is associated with the higher volatility material producing increased the mechanism for this associated with the higher volatility material producing increased tension tension within filmleads which to transverse compression in thefilm. cured film. This compression within the filmthe which toleads transverse compression in the cured This compression leads to leads to greater inter particle contact, allowing easier charge transfer across the film. The result has greater inter particle contact, allowing easier charge transfer across the film. The result has implications implications for volume manufacturing, for a given requirement conductivity (in requirement (in terms Ω/cm) for volume manufacturing, as for a givenasconductivity terms of Ω/cm) the of optimum the optimum processing characteristics can be calculated. Assuming a known production rate a processing characteristics can be calculated. Assuming a known production rate and a dryingand time drying time required to reach within 5% of the final line resistance, then the relative merit of required to reach within 5% of the final line resistance, then the relative merit of multiple thin layers as multiple thin compared to abe single thick layer can be established. Thisout analysis was carried compared to alayers singleasthick layer can established. This analysis was carried for a printing time out for a printing time of 30 s per sheet (20 s for the loading in register and 5 s each for the printing of 30 s per sheet (20 s for the loading in register and 5 s each for the printing and unloading) and a and unloading) and conductivity, a drying timewhich to 95% conductivity, which are integer multiple ofof a standard drying time to 95% are an integer multiple of aan standard dryer pass 150 s, i.e., dryer pass of 150 s, i.e., if three passes were required to achieve conductivity which is 95% ofthen the if three passes were required to achieve conductivity which is 95% of the stabilised resistance, stabilised resistance, then the drying time is 450 s (3 × 150 s). The impact of the solvent is clearly the drying time is 450 s (3 ˆ 150 s). The impact of the solvent is clearly shown in Figure 10 where this shown Figure 10 where analysis hasmost beenconductive carried out for theThe least most conductive analysisinhas been carried outthis for the least and material. netand effect of the solvent is material. The net effect of the solvent is to lower the possible line resistance and to reduce the time to lower the possible line resistance and to reduce the time required in order to achieve a given line required in order to achieve a given line resistance, e.g., a 5Ω/cm line could be a achieved by three resistance, e.g., a 5Ω/cm line could be a achieved by three layers through a 90/48 mesh in around layers through a 90/48 mesh in material, around 1000 withleast theconductive most conductive while 1000 s with the most conductive whiles the would material, take around 1400the s. least conductive would take around 1400 s.
9 of 10
Biosensors 2016, 6, 30
9 of 10
4500
4500
4000
4000
3500
3500
3000
32/100
2500
61/64
2000
77/48
1500
90/48
1000
110/34
500 0
Processing Time (s)
Processing Time (s)
Biosensors 2016, 6, 30
3000
32/100
2500
61/64
2000
77/48
1500
90/48
1000
110/34
500 0
0
5
10
15
20
25
30
Line resistance (Ω /cm)
(a)
0
5
10
15
20
25
30
Line resistance (Ω /cm)
(b)
Figure 10. 10. Simulated Simulatedprocessing processingtime/line time/lineresistance resistancefor fora 700 a 700 wide printed for solvent (II) Figure µmμm wide lineline printed for solvent (II) and and for (III)1–5 forlayers 1–5 layers of material. (a) Solvent II—Least conductive; (b) Solvent III—Most conductive. (III) of material. (a) Solvent II—Least conductive; (b) Solvent III—Most conductive.
Each data point on every curve represents the number of layers which are printed, with a single Each data point on every curve represents the number of layers which are printed, with a single layer at the highest resistance value and additional layers following the curve to lower line layer at the highest resistance value and additional layers following the curve to lower line resistances. resistances. The analysis also illustrates that in all instances it is beneficial (from a processing time The analysis also illustrates that in all instances it is beneficial (from a processing time point of view) point of view) to deposit multiple layers instead of a single thick layer. As the thinner layers dry to deposit multiple layers instead of a single thick layer. As the thinner layers dry quickly, then quickly, then the largest time element of the manufacturing process is reduced. Clearly there are the largest time element of the manufacturing process is reduced. Clearly there are simplifications simplifications in this analysis and the choice of processing time as the sole metric. The in this analysis and the choice of processing time as the sole metric. The simplifications include; simplifications include; drying times can only be a multiple of a single pass, i.e., the drying speed drying times can only be a multiple of a single pass, i.e., the drying speed cannot be varied, and cannot be varied, and the assumption that individual layers combine without any physical the assumption that individual layers combine without any physical interaction. In terms of metric interaction. In terms of metric there is also a cost metric, e.g., there will be staff cost element for there is also a cost metric, e.g., there will be staff cost element for printing, while none is required for printing, while none is required for drying. In addition, engineering issues such as additional screen drying. In addition, engineering as additional screen wear, increased “drying in” with wear, increased “drying in” withissues finer such meshes or difficulties in interlayer register may mean thatfiner the meshes or difficultiesscenario in interlayer register mean that by thethe ideal manufacturing scenario differs ideal manufacturing differs from may that predicted analysis. A more complex model from that predicted by the analysis. A Notwithstanding more complex model factorsanalysis into account. should take these factors into account. theseshould issues,take this these processing based Notwithstanding these issues, this processing analysis based on experimental results has demonstrated on experimental results has demonstrated the principal that sequential multi-layer printing provides the principal that sequential multi-layer printing a more time efficient manufacturing method a more time efficient manufacturing method forprovides carbon electrode sensors. for carbon electrode sensors. The primary metric used to characterise the materials during this study has been the sheet and The primary metrictoused characterise the materials during this study been be thecarried sheet and line resistance. In order fullytolink this to sensor performance, further studieshas should out line resistance. In order to fully link this to sensor performance, further studies should be using the materials within a sensor structure. The work completed is equally applicable tocarried other out using the materials a sensor structure. The work completed equallyPV applicable other industries such as low within temperature electrical circuits and the nascentisflexible industrytowhere industries such as low temperature electrical circuits[16]. and the nascent flexible PV industry where carbon carbon is being proposed as an electrode material, is being proposed as an electrode material, [16]. 5. Conclusions 5. Conclusions An experimental investigation on the effect of solvent on the performance of screen printed An experimental investigation on the effect of solvent on the performance of screen printed carbon materials and their process-ability has been undertaken. The study has identified that lower carbon materials and their process-ability has been undertaken. The study has identified that lower boiling point solvents can lead to more conductive films being produced, with significantly boiling point solvents can lead to more conductive films being produced, with significantly improved improved manufacturing throughput and no identifiable detrimental effect on process stability. It manufacturing throughput and no identifiable detrimental effect on process stability. It has also has also proposed a multi-layer approach to film deposition is a more time efficient means of proposed a multi-layer approach to film deposition is a more time efficient means of producing high producing high conductivity carbon paste films. conductivity carbon paste films. Acknowledgments: The authors would like to thank EPSRC, Innovate UK, Welsh Assembly Government and Acknowledgments: The authors would to thank the SPECIFIC IKC where this work waslike carried out. EPSRC, Innovate UK, Welsh Assembly Government and the SPECIFIC IKC where this work was carried out. Author Contributions: B.P. and E.J. conceived and designed the experiments; B.P., E.J. and P.G. performed the Author Contributions: B.P. and E.J. conceived and designed the experiments; B.P., E.J. and P.G. performed the experiments and and analyzed analyzed the the data; data; E.J. E.J. wrote wrote the the paper. paper. experiments Conflicts of of Interest: Interest: The The authors authors declare declare no no conflict conflictof ofinterest. interest. Conflicts
Abbreviations The following abbreviations are used in this manuscript:
Biosensors 2016, 6, 30
10 of 10
Abbreviations The following abbreviations are used in this manuscript: BP
Boiling point
References 1.
2. 3. 4.
5. 6. 7. 8. 9.
10. 11. 12.
13. 14.
15. 16.
Švancara, I.; Vytˇras, K.; Kalcher, K.; Walcarius, A.; Wang, J. Carbon paste electrodes in facts, numbers, and notes: A review on the occasion of the 50-years jubilee of carbon paste in electrochemistry and electroanalysis. Electroanalysis 2009, 21, 7–28. [CrossRef] Švancara, I.; Walcarius, A.; Kalcher, K.; Vytˇras, K. Carbon paste electrodes in the new millennium. Cent. Eur. J. Chem. 2009, 7, 598–656. [CrossRef] Renedo, O.D.; Alonso-Lomillo, M.A.; Martínez, M.J.A. Recent developments in the field of screen-printed electrodes and their related applications. Talanta 2007, 73, 202–319. [CrossRef] [PubMed] Hart, J.P.; Crew, A.; Crouch, E.; Honeychurch, K.C.; Pemberton, R.M. Some Recent Designs and Developments of Screen-Printed Carbon Electrochemical Sensors/Biosensors for Biomedical, Environmental, and Industrial Analyses. Anal. Lett. 2005, 37, 789–830. [CrossRef] Wang, J.; Tian, B.; Nascimento, B.; Angnes, L. Performance of screen-printed carbon electrodes fabricated from different carbon inks. Science 1998, 43, 3459–3465. [CrossRef] Morrin, A.; Killard, A.J.; Smyth, M.R. Electrochemical characterization of commercial and home-made Screen-Printed Carbon electrodes. Anal. Lett. 2003, 36, 2021–2039. [CrossRef] Kaempgen, M.; Roth, S. Transparent and flexible carbon nanotube/polyaniline pH sensors. J. Electroanal. Chem. 2006, 586, 72–76. [CrossRef] Park, H.K.; Kim, S.M.; Lee, J.S.; Park, J.-H.; Hong, Y.-K.; Hong, C.H.; Kim, K.K. Flexible plane heater: Graphite and carbon nanotube hybrid nanocomposite. Synth. Met. 2015, 203, 127–134. [CrossRef] Lei, H.; Pitt, W.G.; McGrath, L.K.; Ho, C.K. Resistivity measurements of carbon-polymer composites in chemical sensors: Impact of carbon concentration and geometry. Sens. Actuators B Chem. 2004, 101, 122–132. [CrossRef] Jewell, E.H.; Claypole, T.C.; Gethin, D.T. Viscosity control in the screen printing of ceramic transfers. Surf. Coat. Int. Part B Coat. Trans. 2003, 86, 155–163. [CrossRef] Jewell, E.H.; Claypole, T.C.; Gethin, D.T. The effect of exposure to inks and solvents on squeegee performance. Surf. Coat. Int. Part B Coat. Trans. 2004, 87, 253–260. [CrossRef] Phillips, C.; Beynon, D.; Hamblyn, S.; Davies, G.; Gethin, D.; Claypole, T. A Study of the Abrasion of Squeegees Used in Screen Printing and Its Effect on Performance with Application in Printed Electronics. Coatings 2014, 4, 356–379. [CrossRef] Holmes, D.M.M.; Tegeler, F.; Clegg, W.J.J. Stresses and strains in colloidal films during lateral drying. J. Eur. Ceram. Soc. 2008, 28, 1381–1387. [CrossRef] Katsumata, R.; Ata, S.; Kuboyama, K.; Ougizawa, T. Evaporation rate effect on starting point of shrinkage stress development during drying process in solvent cast polymer film. J. Appl. Polym. Sci. 2013, 128, 60–65. [CrossRef] Smits, F.M. Measurement of sheet resistivities with the four point prose. Bell Syst. Tech. J. 1958, 37, 711–718. [CrossRef] Murakami, T.N.; Ito, S.; Wang, Q.; Nazeeruddin, M.K.; Bessho, T.; Cesar, I.; Liska, P.; Humphry-Baker, R.; Comte, P.; Péchy, P.; et al. Highly Efficient Dye-Sensitized Solar Cells Based on Carbon Black Counter Electrodes. J. Electrochem. Soc. 2006, 153, A2255. [CrossRef] © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).
