Full solar energy is 1350 watts/square metre.
Solar cells are, at best, 50% efficient
so that's 675 watts/sq metre
1 horsepower is about 750 watts.

So, in very rough terms, one horsepower per square metre.

A modern diesel loco generates 5,000 to 10,000 horsepower, an automobile 100 to 150.

Clearly, this is peak power, used when accelerating or climbing a gradient. Normal output is about 25% of peak.

so you are going to need a strip of PV cells 1 metre wide and angled directly at the sun or 2 metres wide on the train roof, and 1.25 to 2.5 km long to power this loco.

In practice, most freight trains are multi-headed. For example, the record-setting Iron ore hauls in Western Australia (82,000 tonnes load, 7.35 km in length) have eight 6000 hp engines, and the track is laid with very slight gradients!.

Basically, this would take an amount of PV cells 4 times the available length of the train to give all that peak HP, or the exact length of the train to give the (guess) of 25% for cruising. So it's marginally possible.

There are a bunch of design issues - how do you top-load crushed ore into hopper wagons if you use the roof space for PV cells? (obviously, you have a lid/roof with the cells on that closes, and if you have this, you might as well have a sun-pointing mechanism). How do you protect fragile glass-based PV cells from falling rocks?. How do you supply the extra power for acceleration (e.g. 6 of the 8 locos are "normal" ones for the acceleration phase). Do you try to capture the massive energy of this load running downgrade or under braking with a hybrid (battery) system?

It *could* work, if PV cells become cheap enough and diesel fuel expensive enough. (But I think it would still be cheaper to electrify the railway, use fully electric locos, and have a power station running the electric lines, and using multiple trains on one track to share the input/output of gradients and acceleration/braking)