Professional Documents
Culture Documents
Incorporation of Tin On Copper Clad Laminate To Increase The Interface Adhesion For Signal Loss Reduction of High-Frequency PCB Lamination
Incorporation of Tin On Copper Clad Laminate To Increase The Interface Adhesion For Signal Loss Reduction of High-Frequency PCB Lamination
PII: S0169-4332(17)31373-9
DOI: http://dx.doi.org/doi:10.1016/j.apsusc.2017.05.061
Reference: APSUSC 35996
Please cite this article as: Chong Wang, Na Wen, Guoyun Zhou, Shouxu Wang, Wei He,
Xinhong Su, Yongsuan Hu, Incorporation of Tin on Copper Clad Laminate to Increase
the Interface Adhesion for Signal Loss Reduction of High-frequency PCB Lamination,
Applied Surface Sciencehttp://dx.doi.org/10.1016/j.apsusc.2017.05.061
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
Incorporation of Tin on Copper Clad Laminate to Increase the
PCB Lamination
Chong Wang1,2, Na Wen1, Guoyun Zhou1, Shouxu Wang1, Wei He1,2*, Xinhong Su3, Yongsuan Hu3
(1State Key laboratory of Electronic Films and Integrated Devices, University of Electronic Science and
17
Ga
NEAP
NEAP Brown Oxide
Copper
Surface
Insert
Loss of
Highlights:
1. We substitute the copper surface of PCB by tin to decrease the IEP (isoelectric point) and to
increase the peel strength of copper foil with prepregs, which is called as NEAP (Non-etching
adhesion process).
2. The NEAP creates a higher peel strength with smooth surface in compatibility with high
frequency signal transmission, compared with the traditional brown or black oxidation method by
copper etching.
18
3. The actual composition of tin to increase the peel strength was studied in mechanism.
Abstract. A novel method of improving the adhesion between copper and prepreg in high frequency
PCB was proposed and studied in this work. This process which aimed to decrease the IEP
(isoelectric point) of the copper to obtain higher adhesion, was achieved by depositing a thin tin layer
with lower IEP on copper. It was characterized by scanning electron microscopy (SEM), 3D
microscope, peel strength test, X-Ray thickness test, grazing incidence X-ray diffraction (GXRD),
X-ray photoelectron spectroscopy (XPS), Agilent vector network analyzer (VNA), which confirmed
its excellent adhesion performance and outstanding electrical properties in high-frequency signal
transmission compared with traditional brown oxide method. Moreover, the mechanism of achieving
(Isoelectric Point)
1. INTRODUCTION
High-frequency signal transmission needs the formation of fine copper lines in high-density
including brown oxide or black oxide creating a rough copper surface provides the essential
mechanical bonding to prepreg with its micro-anchor to ensure excellent thermal reliability for
PCB.[2][3] It is believed that this technique will eventually encounter obsoletes in the future PCB for
the applications with high frequency over 25 GHz or in THz, including the next 5G communication,
GPU (Graphics Processing Unit), and THz communication under the high signal transmission loss.
In high-frequency, an AC signal primarily flows nearer to the outside surfaces of a copper conductor
rather than through the entire cross section of the conductor, i.e., a phenomenon called as skin effect.
As frequency increases, the skin effect becomes more pronounced.[4][5] A. Zee and R. Massey
19
conducted a simulation of signal loss, and obtained a result that the signal loss of track treated with
etch-based treatment was 14% more than smooth track, actually the actual measurement showed
more. A. Deutsch also pointed out the dedication of surface roughness to signal integrity was
significant. [6][7] For this consideration, micro-etched or roughened surface is destined to cause signal
loss, resulting in signal integrity transmission problems. Moreover, smoother surface for signal
transmission is required for high resolution ultra-fine lines towards 8 μm in the near future
technology for PCB. A series of problems including low adhesion with substrate, easy presence of
short or open circuits are inevitable, in case of the roughness up to 2~3μm under brown oxidation or
black oxidation method for these ultra-fine lines.[8] Particularly, high frequency signal suffers a great
loss in the long threads signal transmission. Consequently, researchers for high frequency PCB are
turning their eyes to making smooth surface conductors with maintained adhesion between the
Generally, copper and prepreg bonding enhancement should be provided through mechanical
bonding or chemical bonding or both.[9]To maintain a smooth and consistent track surface, the
micro-etch process for mechanical bonding, such as brown and black oxidation methods, is prompted
chemical interaction between metal and polymer is described by the interaction of a metal oxide with
a polymer, which is typified as acid-base chemistry. [11][12]To predict the probability of the interaction
to occur, equilibrium constants for acid-based interactions can be expressed in terms of the
strength of the acid K for polymer and the IEP (Isoelectric Point) for metal, which is indicated as:
The with high value predominates the strong interaction between polymer and metal surfaces,
resulting in strong coordinative covalent bonds between the adherend and adhesive. As shown in
equ.(1), while the adsorption of hydroxide ions on a metal surface covered by a thin oxide layer,
metal with lower IEP should effectively achieve higher bonding strength with a fixed polymer
20
through Van der Waals' force and the hydrogen bond.[13]
The goal of the work is to find a metal conversion of low IEP to enhance the bonding strength
between copper laminate and prepreg. Tin is an excellent metal with low IEP of 4 and other
advantages such as low-cost, easy fabrication, and sufficient compatibility with PCB
production.[14][15] Tin was deposited on copper laminate by electroless process in order to replace the
copper surface (IEP=9) with lower IEP surface.[16] Significant improvement for the adhesion strength
between the copper and prepreg was found. Accordingly, PCB achieved a superior adhesive and
exceptional thermal reliability by tin surface conversion without changing the track profile. Both
signal transmission integrity and fine line fabrication benefited from this surface treatment method.
2. EXPERIMENTS
Process of Copper surface conversion. Copper clad laminates with copper foils of 18 μm in
thickness were used to bond with pregreg. The schematic diagrams of NEAP process on copper
surface were illustrated in Fig. 1. In a typical treatment process, the organic contaminations such as
dry film resist and oily fingerprints were removed in an alkaline cleaning solution (NaOH 60g/L,
NaCO3 30g/L, Na3PO420g/L). It then was followed by the removal of other heavy oxides,
anti-tarnish layers or detergents by rinsing in the acid cleaning solution (5% H2SO4 solution that
added 5% concentrated degreasing solution). The copper clad laminate should be washed at least 3
times in deionized water after each step above-mentioned. Subsequently, the surface of laminate was
coated with a very thin tin by soaking in the tin eletroless-plating solution (SnCl2•2H2O 20g/L,
SC(NH2)2 75g/L, HCl 80g/L, NaH2PO2•H2O 30g/L, surfactant 1g/L and stabilizer 1g/L). The tin
reduction time was controlled from 40s to 400s for its affinity to the plated-tin thickness. Activating
process is necessary for converting tin metal by the mixing of NaNO2 solution with dilute alkali lye
Characterizations. The peel strength of tin-coated copper with pregreg (PI pregreg) was carried out
after the lamination under IPC-TM650 standard at a peeling rate of 10 mm/min with a peel tester,
21
[17]
PT-800. To evaluate the effectiveness of NEAP, copper micro-etched using brown oxidation was
measured in peel strength by the same process above. The copper surfaces processed by NEAP and
brown oxidation were characterized by scanning electron microscopy (SEM, JEOL JSM-6060) for
measurement, respectively. The plated thickness of tin on copper for NEAP was inspected by X-Ray
thickness tester (Fisherscope X-Ray XDLM) and the tin structure before lamination and after
lamination were analyzed by grazing incidence X-ray diffraction (DX-2700, Dandong fangyuan
instrument Co., LTD) with Cu Kα radiation (Bede, D1 System, λ=1.5418 × 10−10 m). X-ray
photoelectron spectroscopy (XPS, ESCA2000, VG Microtech) was designed to detect the chemical
compositions of the separated sides after peeling test for copper and PI, determining the broken
positions in peeling process. Furthermore, the high frequency signal loss attenuation was performed
by taking S-parameter measurements in the range of 1 MHz~20 GHz using an Agilent vector
network analyzer (Agilent ENA Series Network Analyzer E5071C). The test structure vehicle is
presented in Fig.2, in which high frequency laminated copper clad (Panasonic M6) was employed as
Surface characterizations.
The obtained surface morphologies for copper foil after NEAP process was characterized by SEM,
as shown in Fig. 3a. For comparison, copper foil processed with brown oxidation method was
presented in Fig. 3b. The copper surface etched by brown oxidation solution was clearly observed
with roughness of 2~3 μm in horizontal direction, which is required to enhance the adhesion by
physical interlocking mechanism in PCB lamination process. However, this rough characterization
was substituted by a smooth plated tin thin-film on copper after NEAP treatment. As observed on the
SEM, Corresponding 3D microscope images were taken to demonstrate their vertical patterns and the
surface roughness. The typical rough profile of copper foil processed by brown oxidation was
22
revealed, as shown in Fig. 3d, to have ridges with dimensions on the order of 2~3 μm in width by
approximately 2~3 μm in depth. This feature was also affiliated by some smaller irregular shapes,
sizes, and separations. It contributed the roughness in vertical direction producing a roughness of
0.479um in the parameter Ra. By contrast, the sample with NEAP obviously keeps a relatively
smooth surface (Fig. 3c), in consistent with the image observed from SEM in Fig. 3a. Accordingly,
the roughness Ra was calculated to be 0.236 by nanoscope software. It demonstrated that when
copper lad with lines was used as basic substrate, the technique of NEAP shows an excellent
feasibility without any apparent morphology defect, indicating the dominant superiority in
Insertion Loss
It was known that “skin effect” is more pronounced as the high-frequency signal increases,
leading to a severer signal loss over 1 GHz. The insertion loss (S parameter) to clarify the signal loss
of conductive lines treatment by NEAP and brown oxidation was shown in Fig. 4. This is no obvious
difference of signal loss for these two methods for the loaded frequency lower than 5 GHz. Higher
signal loss differences of 1.4 dB/m and 4.23 dB/m at the frequency of 10 GHz and 20 GHz
respectively, was contributed by brown oxidation in comparison with NEAP. The obvious
discrepancy in signal loss can be accounted for the surface roughness differences, in which the
roughened surface for brown oxidation produce higher signal loss due to “skin effect”. Accordingly,
the copper foil treated by NEAP to fabricate conductive lines can effectively reduce the signal loss,
SEM images in Fig. 5 showed the surface morphologies of copper foil at varied immersion
times of 0s, 40s, 80s and 120s in the tin replacement solution. The common low surface profile was
observed at the bare copper of FR-4 copper clad sample in Fig. 5a. After immersed for 40 s, the
tin-plated copper showed a global homogeneous copper surface duplication (Fig. 5b). It is reported
23
that the tin-reduction reaction on copper surface consists of two stages, i.e., the replacement of
copper by tin and the continuous growth of tin by tin self-reduction.[18] The surface morphology
demonstrated a typical tin replacement reaction at the first 40 s. At the continuous immersion to 80 s,
the tin crystalline grain is very notable, indicating the tin replacement reduction was finished and the
self-reduction process has begun (Fig. 5c). When the sample was immersed another 40 s in tin
solution, the lager-sized tin crystalline grain was formed, as shown in Fig 5d.
The thickness of as-plated tin on copper foil varied with depositing time was shown in Fig. 6a. The
time was in the range of 0~200 s. The thickness of the tin layer is only around 0.16μm at duration of
40 s, whereas the immersion time of 200 s increased the thickness to 0.34μm, indicating a thicker
coating with more displacement time. It is notable the curve in Fig. 6a is steeper at the early time of
deposition. This implies that the deposition rate is relatively high during the initial stage and
The corresponding peel strength in Fig. 6b showed the relation of adhesion with the tin thickness
increasing. The tin deposit on copper indicated an obvious improvement in comparison with bare
copper. The peel strength of bare copper foil is 0.43 N/mm. However, the peel strength of a finished
PCB should be higher than 0.5 N/mm according to the IPC standard. After immersed for 80 s, this
value is increased to 0.71 N/mm. The highest peel strength was test to be shown at 0.89 N/mm,
which was contributed by the as-plated tin with thickness around 250 nm or with immersion time of
120 s. At this specific point, the tin crystalline grains begin to be formed as shown in Fig. 5d. If the
crystalline grains continue to grow bigger when the immersion time was prolonged, the adhesion
between copper and prepreg was lowered down. Therefore, to optimize the adhesion for copper with
In the case of solid–liquid (or prepreg) contacts, their interface adhesion has been usually
attributed to surface roughness or to chemical interaction.[19] The peel strengths between copper foil
and prepreg without treatment and treated by NEAP and brown oxidation method were listed in Fig.
24
6c., copper foil treatment is definitely requisite for obtaining high reliability of PCB lamination. For
both the NEAP and brown oxidation, the adhesive strength has demonstrable improvement (almost 2
times) in comparison with that without treatment. The NEAP produces a highest adhesion around
0.89 N/mm at the smooth copper foil plated a very thin tin film. NEAP slightly wins out over the
brown oxidation method of producing a chaotic surface with high roughness for adhesive strength.
Obviously, NEAP achieved this excellent effectiveness by the chemical interaction of as-plated tin
Adhesion interface
The as-laminated copper with PI by NEAP included several peeling interface, such as
copper-copper, copper-tin, tin-tin, tin-PI and PI-PI. The excellent adhesion treated by NEAP was
expected to perform at the interfaces of copper-tin, and tin-PI, i.e., the peel position occurs at the
Fig. 7 showed the optical and SEM images of the separated interfaces after peeling test. It is
observed that a delaminated thin-film attached on tin-plated copper foil. To demonstrate the adhesion
failure, Fig. 8 displays the XPS spectra of the separated PI part and copper part, respectively. The
spectra for PI part, as shown in Fig. 8, indicates the simple detection of elements of C and O, which
are the dominant components of PI. Corresponding elements such as Sn and Cu were not responded
in the XPS analysis. Moreover, the elements of Sn and Cu, which are dominant components of
copper part, did not provide the spectral responses in the XPS data either (Fig. 8). The detectable C
and O suggested that the PI was attached on the surface of tin-plated copper after peeling test, which
was observed in the aforementioned SEM images (Fig. 7b). Therefore, it can be concluded that the
delaminated position for NEAP sample was located at the interface of PI-PI. The delaminating
phenomenon was also verified by brown oxidation method. This showed NEAP like brown oxidation
can produce an excellent bonding for copper and PI prepreg, which outperforms the PI materials
body.
25
Potential Mechanism
After NEAP process, the copper foils before and after laminated with PI as well as after
activated were analyzed by grazing incidence XRD with the incident angle of 1.5°. The results are
presented in Fig. 9. As shown in Fig. 9(a), the activation solution turned the tin into its oxides such as
SnO and SnO2, in which the tin position was not detected. To response the peel strength variation in
Fig. 6b, O content was analyzed by XPS (Fig. 9b). The spectra showed a clear correlation between
the peel strength and O content, in which high O content in Tin layer leads to high peel strength [20-28]
For that, the IEP of SnO and SnO2 was determined by measuring the zeta potential by means of the
electrophoresis and electro-acoustic methods. The IEPs of SnO and SnO2 calculated after
[7]
measurement were 6.1 and 4.3, respectively, where CuO was around 9.2. Therefore, It is
conducted that the peel strength enhancement was resulted from the very thin superficial metal
oxidation. In addition, features of CuxSny in the XRD indicates the deposited tin layer covered on
4. CONCLUSIONS
To achieve smooth copper foil surface bonding to prepreg for PCB interconnection, a new
surface treating method, called NEAP, was introduced by IEP mechanism. Due to the chemical
interaction, the sample of NEAP achieves higher peel strength of 0.89N/mm, whereas that of the
brown oxidation is 0.78N/mm under physical interlocking. This peel strength of 0.89N/mm was
contributed by the as-plated tin with thickness around 250 nm or with immersion time of 120 s, in
which longer or shorter immersion should decrease the peel strength. In fact, the peel strength
enhancement was attributed to high quantity of oxide tin. The peeling interface, verified by XPS, was
happened at the PI-PI, indicating the NEAP method yields a very high bonding force and hence can
replace the brown oxidation method. Besides that, the NEAP showed additional superiority in
ACKNOWLEDGEMENT
26
This work is supported by Guangdong Innovative Research Team Program
(NO.201301C0105324342), the Open Foundation of State Key Laboratory of Electronic Thin Films
(2015B010127012).
REFERENCES
[1] Y. Chen, W. He, X. Chen, C. Wang, Z. Tao, S. Wang, G. Zhou, M. T. Mohamed, Plating
uniformity of bottom-up copper pillars and patterns for IC substrates with additive-assisted
[2] M. Lebbai, J. K. Kim, W. K. Szeto, M. M. Yuen, P. Tong, Optimization of black oxide coating
[3] C. Wang, L. Xiang, Y. Chen, S. Wang, D. Xiao, W. He, Study on brown oxidation process with
imidazole group, mercapto group and heterocyclic compounds in printed circuit board industry,J.
Attenuation,IPC, 3 (2007).
Serino, High-speed signal propagation on lossy transmission lines, IBM Journal of Research and
[6] A. Zee, R. Massey. Advantage of non-etching adhesion promoter on high frequency signal loss,
[7] A. Zee, R. Massey, H. Reischer, Impact of surface treatment on high frequency signal loss
27
Adhesion Promotion System for Dry Film Applications, IMPACT, 3 (2008) 407-410.
Adhesion and Durability Performance of Zirconium Based Conversion Coating on AA5182 for the
adsorption and acid-base interaction behavior. Progress in organic coatings, 28 (1996) 307-311.
[11] G A. Parks, The isoelectric points of solid oxides, solid hydroxides, and aqueous hydroxo
[12] F. M. Fowkes, Role of acid-base interfacial bonding in adhesion, J. Adhesion Science and
isoelectric points of metals and metallic alloys by adhesion of latex particles, Journal of colloid and
[14] M. Kosmulski, Compilation of PZC and IEP of sparingly soluble metal oxides and
hydroxides from literature, Advances in colloid and interface science, 152 (2009) 14-25.
[16] N. Kallay, Z. Torbic, M. Golic, E. Matijevic, Determination of the isoelectric points of several
[17] Y. Wei, J. W. Hutchinson, Interface strength, work of adhesion and plasticity in the peel test,
[18] Y. Fujiwara, Sn deposition onto Cu and alloy layer growth by a contact immersion process,
and interface chemical reactions of Cu/polyimide and Cu/TiN by XPS, Applied Surface Science, 205
28
(2003) 128-136.
[20] E. P. Plueddemann, Nature of Adhesion Through Silane Coupling Agents, Silane Coupling
[21] D. Zhu, X. Nai, S. Lan, S. Bian, X. Liu, W. Li, Surface modification of magnesium hydroxide
sulfate hydrate whiskers using a silane coupling agent by dry process, Applied Surface Science, 390
(2016) 25-30.
[22] C. C. Wang, J. Y. Ying, Sol-gel synthesis and hydrothermal processing of anatase and rutile
[23] K. S. Khoo, E. J. Teh, Y. K. Leong, B. C. Ong, Hydrogen bonding and interparticle forces in
platelet α-Al2O3 dispersions: yield stress and zeta potential, Langmuir, 25 (2009) 3418-3424.
[24] E. J. Sadler, A. C. Vecere, Silane treatment of mineral fillers: practical aspect, Plastics rubber
[25] E. P. Plueddemann, Silane primers for epoxy adhesives, J. adhesion science and technology, 2
(1988) 179-188.
improved adhesion: a critical review, J. Adhesion Science and Technology, 7 (1993) 1091-1127.
of elastomers in the rubbery plateau by carbon black, Kautschuk und Gummi, Kunststoffe, 47 (1994)
349-353.
[28] J. C. Bolger, Acid base interactions between oxide surfaces and polar organic compounds,
[29] K. N. Tu, Irreversible processes of spontaneous whisker growth in bimetallic Cu-Sn thin-film
on copper circuitries in circuit boards, Surface and Coatings Technology, 202 (2007) 417-424.
29
30
CAPTIONS:
Fig. 1. Schematic diagrams illustrating the process of NEAP on copper foil surface.
Fig. 3. (a) SEM surface morphology of copper foil by NEAP; (b) SEM surface morphology of
copper foil by brown oxidation; (c) 3D microscope morphology of copper foil by NEAP; (d) 3D
Fig. 5. SEM images of copper surface morphologies at varied immersion times: (a) 0 s, (b) 40 s, (c)
Fig. 6. The thickness (a) and the peel strength (b) of tin layer varied with immersion time(c)
Fig. 8. The X-ray photoelectron spectroscopy (XPS) results of the PI part(a) and the metal part(b) ((c)
Fig. 9. (a) Graizing incidence X-ray diffraction patterns of copper foils after oxidation in NEAP
process (b) XPS about the content of O on the surface after different immersion time
31
Fig. 1
32
Fig. 2
33
Fig. 3
34
Fig. 4
0
NEAP
Brown Oxide
-15
Signal Loss(dB/m)
-30
-45
-60
-75
0 5 10 15 20
Frequency(GHz)
35
Fig. 5
36
Fig. 6
(a)
0.4
(b) 0.9
0.3 0.8
Peel Strength(N/mm)
Tin Thickness(μm)
0.2 0.7
0.6
0.1
0.5
0.0
(c) 1.0
0.8
Peel Strength(N/mm)
0.6
0.4
0.2
0.0
NEAP Brown Oxide Copper
37
Fig. 7
38
Fig. 8
(a) Atomic %
O1s (b) Atomic %
O1s
C1s
C1s C1s 70.99
C1s 69.97
O1s 22.9 O1s 22.94
N1s 3.05 N1s 3.08
Si2s 4.08 Si2s 3
Intensity
Intensity
N1s
N1s
Intensity
39
Fig. 9
31.86%
Intensity
4