By placing a foil at one side of the antenna, you have basically made an reflector to distort the radiation pattern and force the energy out of one side of the antenna to make it semi directional, as opposed to the intention of the design, which was to be omni-directional. Doing this however will almost certainly cause the inductance and/or capacitance of the antenna to alter, as well as the impedance (which needs to stay as close to 50 ohms to present an acceptable voltage standing wave ratio to the radio). If the VSWR gets too high for the radio it leads to energy being reflected back to the transceiver, which could damage it. I am not sure if the power levels in consumer 4G radios is that high to begin with (0.5-3W maybe?), but I believe that the radio power output is managed by the receiving cell, so the EIRP will fluctuate as the conditions and loads on the cell dictate. You could take a chance and run it like this, and it might be ok for a while, but there is a risk you could damage the radio in your router, most likely the output stage of the amplifier.
If it was as easy as sticking a bit of foil to one side of a router there would be no antenna manufacturers. Same goes for homebrew designs off the internet imo. Unless you can measure the antenna you have created with an vector or antenna analyzer, you simply are not going to know what you have plugged into your device.
On the plus side, at least you now know u can gain some possible speed by adding an directional antenna.
If you have a genuine interest in making antennas, you can download lots of free design software, like
EZNEC and
MMANA-GAL
You can input the antenna dimensions and run simulations in free space and over modeled ground conditions to see how a design will perform before building anything. This will give you an understanding of the complexities and parameters that need to be considered to produce the best results.