Soldering may be "acceptable", but to me the manufacturer-supplied crimps are "approved". I don't want to drag this on but would like to make 3 points, which I'll expand on as briefly as I can...

1. The O2 sensor absolutely requires a (minute) supply of reference air to generate it's signal voltage.

2. The inherent difficulty of soldering stainless steel (fluxes, heat etc.) potentially risks the sensor's access to reference air via the signal wire.

3. The possibly superior conductivity of a soldered sensor signal wire splice is not needed.

(1) Early sensors got reference air from a hole in the sensor capsule, but not any more. Most sources say the air now comes in via the wire. I accept that because there seems to be no alternative.

(2) Assuming that the air access is through the wire, soldering those stainless steel strands could introduce flux run-off or, in the case of silver soldering, the need for possibly corrosive fluxes and high temperatures.

(3) A simple on-the-car exercise will show that the ECU does not need the potential super conductivity of a soldered splice...

With engine running at a warm idle, disconnect the sensor from the ECU (idle speed may change a little). Hold the ECU terminal with one hand and touch the Battery + terminal with a finger of the other hand.

Your body will carry a weak positive potential from the battery, that ECU "sees" as a very rich signal from the O2 sensor. In response, the ECU will drastically lean the mixture to the point of stalling the engine.
That miniscule + potential all it takes.


--
Bruce Young, '93 940-NA (current), 240s (one V8), 140s, 122s, since '63.