Your four requirements aren't necessarily correct, but rather than explain in detail why each criterion is slightly wrong, I'll make a long rambling post on hypervelocity impacts, because I've done that a bunch of times before on this site and apparently nobody can stop me.
The first, and most important thing to remember is that hypervelocity impacts aren't like normal things smacking into each other. The impact pressures vastly exceed any plausible yield strengths, which means that both the impactor and impactee undergo plastic flow... this means they're more like liquids splashing off each other. This is a complex thing to explain, so you could have a read of this lengthy PhD thesis for more detailed explanations: Penetration of a shaped charge.
Or you can just take my word for it.
Modern shaped-charge weapon research gives us a handy equation (A Jet Penetration Model Incorporating Effects of Compressibility and Target Strength calls it the Hill-Mott-Pack equation, but other names are associated with it too including Birkhoff and Tarantello, but I digress) and it looks like this:
$$P_d = \ell \sqrt{\rho_j \over \rho_t}$$
Where $P_d$ is the depth of penetration, $\ell$ is the length of the penetrator, $\rho_j$ is the density of the penetrator (with j-for-jet, as the authors are thinking of shaped charge jets which are handy present-day hypervelocity military hazards) and $\rho_t$ is the density of the target. (Note: this isn't a perfect model for all hypervelocity impacts, especially against armor plates that can deform instead of massive thick slabs of metal, but it serves to illustrate the problem. Real world penetration depths are likely to be higher!)
If your armor is thinner than $P_d$, then your spacecraft is going to have a bad day. However, big thick shells of dumb armor are heavy, expensive and probably impractical... you're probably not going to put a meter thick shell of tungsten around your space station, for example.
Whipple shields don't remove that all-important $P_d$ term from the above equation, and they're not usually thick enough to substantially reduce the velocity of the impactor.
What they do, is to reduce the $\ell$ term.
As you should already know, Whipple shields are not expected to stop everything. Here's what happens in an impact:
- The impactor hits your Whipple shield.
- It is probably denser than the shield and longer than the shield is thick.
- The impactor and shield act like liquids and splash off each other.
- Most of the impactor will continue on through the shield.
Your Whipple shield needs to be thick enough and dense enough that an impactor won't just blow through it without noticing. On impact, the tip of the impactor and the Whipple shield itself will splash out of the way, and unless the impactor is very small (like bit of gravel) most of it will carry on through.
The energy of the impact, however, will send a shockwave up through the impactor, and at least initially that shockwave will be strong enough that the molecular bonds holding the impactor together will not be able to resist it, eg. the end of impactor will explode. (I don't have any useful figures on the nature of this disruption... the design of Whipple shields shows that it happens, but the details of the why and how much are very hard to come by and might even be classified. Research on the effects of high-density projectiles is limited, and only seems to note that "they go through more armor").
The cloud of debris from the disrupted penetrator will initially still be very dense, and very dangerous, so you need to give it room to expand. How fast it expands depends on too many variables to consider here (the strength and density of the penetrator, the thickness and density of the Whipple shield, the spacing material or lack thereof between layers, the speed of the penetrator, etc etc). This is why Whipple shields require a standoff distance, to let the cloud of debris expand.
Instead of one long deadly penetrator, there are now a number of much smaller penetrators, spread out over a much larger area. The $\ell$ term is therefore much lower, and your armor thickness can therefore be much less.

(Image taken from Micrometeoroid and Orbital Debris Environment & Hypervelocity Shields, showing how important the standoff distance is.)
So, back to your four requirements:
- Lightweight. Everything on a spacecraft needs to be lightweight, really. Bits of the shield will hit the craft in the event of an impact, but they can't be any more dangerous than the penetrator itself even if they're made of something really heavy, because physics doesn't work that way. The outer layer needs to be heavy enough to disrupt expected incoming projectiles.
- Flexible. What you probably mean here is tough and strong. This prevents the hole made by the penetrator from being any bigger than it has to. Steel isn't particularly flexible, but it is quite tough and strong, for example.
- Hard. Doesn't really matter, because at the impact pressures involved everything is behaving like a fluid and fluids can't meaningfully be "hard". You want a certain minimum amount of areal density in order for the penetrator to release enough energy when it hits to disrupt it. This means you can have a thin, high-density shield or a thick, low-density shield.
- Spready. This isn't a property of the shield material... indeed, as it is behaving like a fluid in the impact, it can't really have many useful properties from its shape beyond plain old thickness. The spreadiness of your Whipple shield is a factor of its stand-off distance, eg. the spacing between layers. The more spacing the better, but obviously you are limited by the design of your spacecraft.
Here are some diagrams of real-world shields on the ISS:

Note use honeycomb and corrugated aluminium components, but for intermediate layers where they do a better job at catching hypervelocity crud without adding too much weight, not for outer layer spreadiness, see page 27 of the MMODE&HS presentation). There are two kinds of shield which don't have a solid outer layer at all, using either basalt fibers or woven metal threads. The non-metallic shields use much more spacing for the same protection. Other work mentioned in Shields for Enhanced Protection Against High-Speed Debris talk about
exterior “bumper” layers composed of hybrid fabrics woven from combinations of ceramic fibers and high-density metallic wires or, alternatively, completely metallic outer layers composed of high-strength steel or copper wires. These shields are designed to be light in weight, yet capable of protecting against orbital debris with mass densities up to about 9 g/cm3, without generating damaging secondary debris particles.
So what do we end up with? Well, its a bit boring and closer to old-school wet-navy battleship materials than cutting-age space age magic. Metal sheets work just fine, though if you're really weight-limited then fancier materials, especially in woven form, can also perform well. Aluminium sheet is OK... denser and tougher metals might be better, but research is lacking.
Really, all the clever stuff is in the inner layers... the bits where you have to deal with large volumes of low velocity debris, where being able to absorb energy by deforming or shatter projectiles with hard layers can actually have a useful effect. That's a much harder question to answer, but happily it isn't the one you asked!
Consider also, though:
- Hypervelocity impacts are not the only hazards in space.
- Metal sheets are good at breaking up hypervelocity impactors, but are ineffective against neutral particle radiation (eg. neutrons from nuclear weapons) and downright dangerous against charged particle radiation (due to bremmstrahlung).
- Steel is a lovely easy material to cut with a laser.
A more general purpose outer layer on a military object in space might look more like enriched boron nitride nanotubes (tough, refractory, good neutron absorption cross-section) and UHMWPE to catch charged-particle radiation... but again, this isn't the question you've asked, but is definitely a problem you should be thinking about.
hard science
tag. Perhaps you meant to usescience-based
? $\endgroup$