High-quality flexible printed MRI receive coils towards garment integration
Pierre Balthazar Lechene1, Joe Corea1, Anita Flynn1, Michael Lustig1, and Ana Arias1

1EECS, UC Berkeley, Berkeley, CA, United States

Synopsis

Close proximity of MRI receive coils to the patient can allow an increase of signal-to-noise ratio (SNR). Integrating the coils into garments that tightly conform to the body can provide such proximity. This work develops flexible printed MRI coils on a mesh with the potential to be integrated into garments. The dielectric used in the coil’s capacitors is optimized to provide SNR within 91% of conventional coils. Encapsulation enhances the coils mechanical robustness, allowing bending below 1mm of radius of curvature. It is shown that, by cutting and sewing, the coils can be tailored to intimately fit a brassiere cup.

Introduction

Close proximity of printed flexible MRI receive coils to the patient’s body can allow an increase of SNR in images compared to rigid coils 1,2. Such proximity could be achieved by tightly integrating the coils into garments that can intimately conform to complex body shapes. Printed coils for garment integration would need to demonstrate good imaging quality, flexibility and be compatible with garment manufacturing processes. In previous works, fully printed MRI coils were demonstrated with a focus on the silver inks3. Despite an order of magnitude improvement of the conductivity of the deposited metal, the coils unloaded quality-factor (Q) stagnated at 20, indicating that the dielectric was the main source of losses. While the plastic substrate can be used as the dielectric, printing the dielectric from an ink provides more freedom to optimize the substrate for specific needs. In this work, we investigate the properties of a new dielectric ink to improve the imaging quality and the mechanical flexibility of our printed MRI coils. By printing on a mesh substrate, the coils can then be cut and sewn to conform to complex curved shapes such as a brassiere breast cup.

Methods

Our coils are fabricated by screen-printing three patterned layers: a bottom layer of silver (Dupont 5064H) forms the conductive loop of the coil, a dielectric layer forms the separator of the capacitors and a final layer of silver completes the capacitors. The resonance frequency of the coil is tuned by adjusting the length of the overlapping capacitor sections. Two substrates are used, polyethylene terephtalate (PET) and a mesh of polyether ether ketone (PEEK). For the dielectric, two inks are compared, Creative Materials CM 116-20 which was used previously 3 and poly-dimethyl siloxane (PDMS, Sylgard 184). The dielectrics are compared by measuring their dielectric loss tangent 4. They are included in test coils made with copper foil and then in printed coils. The coils loaded Q-factor is measured on a conductive 0.68 S/m phantom 5. To test the mechanical stability of the silver-PDMS stack, layers of silver on top of PDMS are bent along rods of diminishing radius from 2 cm to 1mm. The resistance of the silver lines is measured before and after bending with a Keithley 2400. A top encapsulation layer is realized by blade-coating PDMS.

Results

PDMS offers an order of magnitude improvement in the loss tangent at 127 MHz over the previous dielectric ink (Table 1). This improvement translates to the copper foil coils which have unloaded Q-factors of 175 for the PDMS compared to 20 for the previous ink. In printed coils, the difference narrows to unloaded Q values of 70 and 17 respectively. This suggests that the conductivity of the silver ink becomes the main source of losses for the PDMS coil, while the dielectric is limiting for the CM 116-20 one. Once loaded, the Q-factors of the coils equal 10.7 with the PDMS one and 6.7 with the previous ink. In scans of a phantom on a Siemens 3T Trio scanner, the PDMS coil provides 91 % of the SNR obtained from a state-of-the-art control coil composed of copper foil and porcelain capacitor (Fig. 1). Coils printed on PET and PEEK-mesh have the same Q-factors in all situation. However, bending tests indicate that the maximum bending radius without any resistivity increase of silver laying on PDMS is 6 mm on PET and 8 mm on PEEK-mesh. Encapsulating the top silver with another layer of PDMS mitigates this failure and allows to bend to radiuses below 1 mm on both substrates without electrical loss. Encapsulation does not significantly modify the Q-factor or resonance frequency of the coils. PEEK-mesh offers the advantage of letting the silver ink permeate through its pores (Fig. 2), thus providing excellent electrical contact between the recto and the verso of the mesh. Two separate traces can be electrically connected by superposing them and stitching them together. Conductive epoxy (Chemtronix CW2400) can also be used to strengthen the contact. An overlap of 5 mm is enough to provide excellent electrical contact and mechanical stability, withstanding bending radius of 1mm without incurring losses. Cutting and sewing can be used to tailor the coils to fit a brassiere breast cup (Fig. 3).

Conclusion and Outlook

By improving the dielectric ink, we were able to print robust flexible coils offering 91% of the SNR of conventional coils. Printing on a PEEK-mesh substrate allows to tailor the coils to fit complex shapes, opening the way for integration of printed coils in garments.

Acknowledgements

This work was made possible with funding from GE, NIH R21 EB015628, NIH R01 EB019241, the Hellman Fellowship, the Okawa Fellowship, the Bakar Fellowshp, and the Sloan Research Fellowship.

References

1.J. Corea et. al, Screen Printed Flexible 2-Channel Receive Coil Array, ISMRM 2012
2.J. Corea et. al, Design and Implementation of Flexible Printed Receive Coils Arrays. ISMRM 2013
3. B. Lechene et al., Materials for printed MRI surface coils: towards better image quality and coil flexibility, ISMRM 2015
4. IPC TM-650 Test Method
5. C. Hayes et al., Noise Performance of Surface Coils for Magnetic-Resonance Imaging at 1.5-T. Med Phys 1985; 12, 604-607

Figures

Table 1: Comparison of the properties of the two dielectric inks, as films and included in copper coils or in printed coils.

Fig.1: MRI scans of a conductive phantom with a printed PDMS based coil (left) and a control conventional coil (right). The printed coil image reaches 91% of the SNR of the control. Scale bar is 10 cm.

Fig.2: Front (left) and back (right) of a PEEK-mesh substrate with silver ink printed on the front side. The ink (red arrows) completely permeates the mesh (green arrows) and reaches its back, providing electrical contact between the two sides.

Fig.3: Silver traces printed on PEEK mesh can be cut and sewn back together with (A) or without (B) conductive epoxy and retain equivalent electrical conductivity. As a result, printed coils that could not fit on a breast cup (C) can be tailored to tightly conform to this shape (D).



Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)
0176