All present techniques using conventional plastics (filled with conductive fillers) or also zinc flame spray coating are commercially unattractive and technically insufficient, especially upon the increasing demand of EMI shielding efficiency in our modern world. This induced a trend back to metal housings in the last years, at least for the medium and high level EMI shielding specifications. Several single exceptions do not contradict to this statement.
A (thermoplastic) polymer blend with ICPs, or a powerful ICP coating would in principle offer to match all demands, as the conductivity could be even throughout the whole mass, in contrast to metal fiber or flake filled plastic compounds (problems in corners and small rips).
Highly conductive polymer blends using polyaniline have been developed by us for EMI shielding purposes [1]. By dispersing polyaniline in a matrix polymer like PVC, PMMA or polyester, conductivity figures of around 20 S/cm, and in some cases up to 100 S/cm, can be achieved. Such values, which are higher than has so far been achieved by incorporating carbon black in polymers, promise a very high standard of EMI shielding. The shielding effect is up to 25 dB higher than with carbon black compounds and lies, depending on the frequency of the electromagnetic interference, in the region of 40 to 75 dB for both near and far field. But still a considerable improvement in mechanical values is needed, however, and - as shown below - preferably conductivity levels that are higher by 1 to 2 orders of magnitude. This is the reason, why we are actually not offering such blends for commercial purposes.
Together with L. Shacklette et. al., we had systematically investigated the extent to which the shielding effect depends on conductivity and coating thickness [16 b, section. 5.4 ref. 15].
In general the shielding effectiveness (SE) of a material is defined as the ratio of the transmitted (P0) to the arriving radiation energy (Pi).
(1)
Useful shielding effectiveness values for commercial applications in electronic housings are in excess of 40 dB at frequencies of 1 GHz. For military applications and for near-field shielding requirements even better performance in the region of 80 - 100 dB is required.
As long as the conductive component is uniformly and well dispersed in the polymer matrix, the shielding effectiveness proves in theory and in practice (fig 2) to be a function of conductivity and thickness. Shacklette and Colaneri have developed a correlation function describing this dependence.
The conductivity of conductive plastic specimen was in the region of 0.1 S/cm to 10 S/cm. The full expression for far-field shielding effectiveness therefore possesses two special limits in the region of the megahertz frequencies. These depend on whether the frequency of the radiation to be shielded is higher or lower than the frequency at which the coating thickness d is equal to the classic surface depth d. The classic surface depth is the depth to which the radiation of frequency w penetrates into the material and thereby undergoes a reduction in intensity to the 1/e -th part of its original strength. This surface depth is described by:
(2) 
where m0 = 4p x 10-7 H/m is the permeability of the charge-free space.
A sample is described as ”electrically thin" if d << d and as ”electrically thick" if d >> d. The crossover frequency wc at which d = d, is determined from equation (5.2) as:
(3) 
For frequencies that are very much lower than wc = 2/ (d*m0*d2), the shielding effectiveness in the case of an ”electrically thin" shielding becomes independent of the frequency, and equation (1) is reduced to:
(4) 
Here Z0 is the impedance of the free space and Rs is the surface resistivity ( = 1/ (s*d). For frequencies higher than wc(d > d) the effectiveness of far-field shielding can be approximated by:
(5) 
where e0 = 107/ (4pc2), the dielectric constant of the free space (c is the speed of light).
The first term in this equation is the contribution to shielding, taking account of the simple reflection wave that impinges on the front and rear of the surface of the envelope. The second term represents the attenuation of this wave by absorption during its passage through the envelope. At high frequencies (w >> wc) the second term dominates and the shielding capacity increases with frequency.
For near-field shielding, just as for far-field shielding, it is possible to approximate a general expression by means of the limits of the electrically thin and electrically thick samples. At the limit of an electrically thick shielding (d/d >> 1 or w >> wc ) near-field shielding effectiveness can be approximated as:
(6) 
The first term on the right-hand side of the equation is interpreted as the shielding due to reflection, and the second as the shielding due to absorption. It is expected that the shielding due to reflection will decrease by about 30 dB per decade of frequency increase (steps of 10). The entire near-field shielding will naturally never be smaller than the far-field shielding. As the wavelength decreases, the near-field shielding zone approaches the far-field shielding.
Here too we find a difference in frequency dependence for electrically thick and electrically thin samples. The approximation d/d << 1 (or w << wc) leads to:
(7) 
Unlike equation (6) this equation shows a decrease of 20 dB per decade of frequency increase and 20 dB per decade of surface resistivity, RS = 1/(s*d). It may therefore be concluded that the most effective near-field screening can be expected in the region of 20 to 30 dB for every decade of frequency increase and 20 dB to 10 dB for every decade of increase in specific resistance.
The fig 2 and 3 show theoretically expected values of shielding-effectiveness (after equations (6) and (7) compared with measured values (fig 3). The highest and therefore best value for specific conductivity was obtained for far-field shielding. At low frequencies the measured curve and the predicted curve agree exactly. At high frequencies, by contrast, the measured shielding efficiency tends to increase with frequency earlier than theoretically predicted. Since this deviation is very small, it may be assumed that the conductivity of this sample increases at higher frequencies. One can in fact expect a frequency-dependent conductivity to occur where a conductive material is dispersed in a matrix. This is in accordance with our finding in work together with Nimtz and Pelster, cf. [24].
It emerged from these theoretical and practical studies that it is possible to achieve very high values for shielding performance by using intrinsically conductive polymers in form of thermoplastic blends. This was made possible by improving the electrical contacts and reducing the zones of poor conductivity in the ICP blends developed, resulting in outstanding electrical conductivity.
However, in view of the blends available today with conductivities of 10 to 100 S/cm (with still insufficient mechanical properties, though), the technically necessary shielding efficiency require wall thickness of 2 - 3 mm. These are technically unacceptable housing wall thickness values. It would be acceptable to have a thickness of 0.5 to 0.8 mm, including mechanical and probably also flame retardant properties. Coatings on insulating plastics would also be acceptable, but they would have to have about 4 orders higher conductivity values compared to actually achievable values. We are far away from such specifications.
It is therefore unavoidable to develop blends with much higher conductivity, higher by one to two orders of magnitude, and which also display mechanical properties of technical polymers. Here, it is interesting to note that we were able to increase the conductivity of a blend to about 50 S/cm starting with a raw material of about 5 S/cm.
List of Figures:
fig 2: Theoretically expected values of far-field and near-field shielding compared with actually measured values.
fig 3: Near-field shielding effectiveness data as theoretically predicted for the ranges > c and < c and as membered in the “two-chamber box”.
