I have raised through about 30 species of Odonata larvae (at least to the stage where tricks gave way to the logistics of gathering enough food).
In my experience the critical stage is the 2nd instar larva - where you can normally expect high mortality. Once however, you have got it right in terms of food and conditions mortality can approach zero for this stage.
You will need to have a range of potential foods in culture.
Some swear by, but I swear at, ciliates - certainly some ciliates are too well
defended by trichocysts.
I use harpactacoid copepods which I established in lab culture over a decade
ago - I cleaned the culture by serial transfer and feed them a few small blades
of boiled grass every few months while in maintenance phase. I split the
cultures and add food about 2-3 weeks before I need them (usually when I
collect the odonate eggs). Even the smallest emerging larva seems able to
handle at least the nauplii.
Use ONLY harpactacoid copepods - cyclopoids include fierce predators which will
chomp your dragons.
Brine shrimp nauplii may also do, but our aquaculturists feel they lack some
essential nutrients (at least for fish), they also tend to be a tiny bit too
big.
Small cladoceran species (esp the benthic forms, the Chydorus-like ones) are a
useful alternative - some species do much better on these than on
harpactacoids.
By mixing and matching (= trial and many errors) the above have worked for all zygopterans, corduliids and libellulids I have tried. The relative sizes and speed of the prey are critical to achieving an appropriate match.
Aeshnids are different. My only successes with raising aeshnids from the eggs is when I have had access to large numbers of hatching chironomid egg masses. Other egg masses which produce a small larva will also work - the key trick is that you will need lots as the prey will soon outgrow the capabilities of the predators. Second-instar aeshnids are enthusiastic about food, but not all that competent. Matching is needed, outside the appropriate size/mobility range and you can forget it. Third-instar larvae are much more able to look after themselves.
Harpactacoids, from nauplii to adults, provide a wide range of prey size
and, from small larvae struggling with a small nauplius to a large larva
stooping to its minimum prey size, cover about the first half of a dragonfly's
development. Once above the size harpactacoids work you will need to switch
food types. Larger cladocerans or mosquito larvae seem good. Tubificids
(available live from many aquarium shops) are a good base diet but there is
suspicion whether things can really develop on a straight diet of tubifex or
any other similar oligochaete. Comments I have had indicate that a diet
supplement is important. Also note tubifids from stores usually come from a
natural source and may have a range of other organisms included - usually no
problem to fish, but potentially a problem to small dragons. Chironomid larvae
are also useful. These come in a range of sizes and many are easily bulk-raised
from egg masses - a 20-litre drum or some-such with half an apple fermenting in
the bottom suits them fine.
A CAUTION: larval Chironominae usually make tubes, once they are 'tubed' they
are hard for the dragons to access (figures from the chironomid's perspective).
You may need to use a stiff brush on the walls and floor of the container every
few days to get them out of their tubes and make them available.
Another CAUTION: larval chironomids will incorporate dragonfly exuviae into
their tubes, so if you are raising animals to get an exuvial series take care
to inspect the container regularly so you beat the chironomids.
For larger larvae you need a good stock of food. In the last instar most corduliids and aeshnids will polish off maybe 200 largish mosquito larvae ... Big aeshnids can be hand fed using say chopped up blowflies (heads are useless) or tiny bits of raw fish/chicken impaled on a needle. 'Tease' the animal with the food and get a defensive strike ... after a few such encounters the larva will catch on and will come to be fed.
Watch rearing temperatures. Some species like it warmer, others show their intolerance by lying on their backs - dead. Even closely related species can have different temperature requirements.
Use water that has been conditioned and allowed to stand for several days (maybe boiled with a few grass leaves then left - takeaway food containers are good for this, oxygen diffuses readily through these sorts of plastic, but bacteria stay out).
For raising individuals use a smallish container (say a 25 x 75mm vial with around 2cm of water). Always place something in the container for the larva to cling to. Newsprint paper seems adequately sterile. Without a perch/refuge many larvae will just walk and walk and walk … Add food using a wide-bore pipette - a narrow-bore pipette gets more suck but may cause mechanical damage. After a week check for moulting by gently swirling the vial in a dark-field type of lighting (back/side lit and look through at a darkish object). Use the lens effect of the curve of the tube. Moults almost glitter as they pass a light-dark boundary. Repeat the inspection daily after this until the animal moults or dies. If the animal is moulting you are on your way. If it is dead you can stop feeding it.
For bulk raising use rectangular takeaway containers, put plenty of substrate (e.g. scrunched up paper towel etc) so animals can separate out and add food culture. Keep some of the larvae in vials to check individual progress. If all the vialed animals die still check the bulk container as a bit of cannibalism or scavenging on a dead or dying sibling may have brought an animal or two through. As the animals grow check and split them when needed. Bigger individuals will often eat smaller ones and separation is a scale thing - a safe density for instar 2 may produce separations well within the sensory range of instar 4 or 5.