Post-Fabrication Technique to Manage Material Variations in 3-D Printed Microstrip Antenna Substrates

To fulfill the growing demand of next-generation wireless devices, additive manufacturing has emerged as the most economical, environmentally friendly, and high-throughput technique for mass manufacturing. At present, however, 3D-printed antenna substrates show 4%–8% variation in relative permittivity
$(\varepsilon _{r})$
compared with typical tolerance values of less than 2% in commercially available substrates. This relatively large tolerance in 3D-printed substrates is attributed to the variations in the quality of the material cassette, extrusion speed, temperature profile, individual layer thickness, printing orientation, infill patterns, and density (i.e., air gaps). Consequently, the resonant frequency
$(f_{r})$
of antennas on 3D-printed substrates may vary batch-to-batch; this is a major issue for narrow band antennas such as microstrip antennas. This study presents a post-fabrication technique capable of compensating for variations in the
$\varepsilon _{r}$
of 3D printed substrates, focusing on microstrip patch antennas (MPAs). The proposed technique, suitable for correcting the
$f_{r}$
of a single microstrip patch antenna, uses either a single blind via or an array of vias in the 3D-printed substrate. The direction and shift in
$f_{r}$
depend on the location and dimensions of the via. Through electromagnetic simulations, a complete set of guidelines has been provided for selecting the number of vias, their locations, and their dimensions for a desired amount of shift in the
$f_{r}$
of the antennas. Furthermore, the shift of the
$f_{r}$
is explained using a proposed improvised transmission-line model of the MPA (incorporating vias). To confirm the proposed technique, two MPAs were fabricated on a 3D-printed substrate and it is shown that their
$f_{r}$
values can be shifted upward or downward by following the proposed guidelines. The maximum shift in
$f_{r}$
, without significantly affecting MPA performance, can reach up to 13%, which is sufficient to manage variations in
$\varepsilon _{r}$
values of 4%–8%.