(From comments to another user.)
I too experienced what you call the bubble gum effect. I called it "gooey gluey paste".
If I may...:
There are three steps needed:
1. don't give it all the hydration at once. If you give it all the water at once, the finer particles or the bran locks up the water, and it will never leave the glue state. However, even with the lower hydration, it will be gummy/gluey, but only temporarily so.
My WW durum (store bought, roller milled, Sher Fiber Wala) is like this:
a) if I give it 85% water up front, it becomes _permanent_ gluey paste. Nothing will then change it to workable _dough_.
b) If I give it 77% water up front, it becomes gluey paste (bubble gum), but in about 3 hours it absorbs the water and becomes workable dough, to which I can add 12% more water in 3 steps of 4% each.
2. Wait 3 to 8 hours. Durum is glassy, glass-like, aka vitreous, which slows water absorption. Its flour is not powdery like wheat, it is glass-like shards. Tiny shards, but not a "powder" like red or white wheat.
3. Add the final water slowly, in 2 or 3 steps, or it will enter permanent glue state again. Add, wait, add, wait, add, wait.
I think you are possibly operating under three misunderstandings:
1. What you are sifting out might not be the bran. The seive only knows the size of the particles, not where they come from. What if the larger particles are the hard glass-like endosperm, and the small particles are the more easily broken down and softer bran?
Suggestion: don't sift, at least for now. Sifting is just adding another variable.
Durum is not the same species as wheat. same genus, different species. NOT just a different variety/strain like red/white or hard/soft. Therefore..... as we learn to use it, all assumptions about how the flour should behave have to be abandonded because it is not "common wheat". It is Triticum Turgidum Durum, not Triticum Aestivum.
Therefore, don't assume that what is retained in the seive is mostly bran, or most of the bran.
In other words: Durum does not and can not mill and break down like red/white wheat because it is not red/white wheat. It is a different species of plant.
2. To get rid of the gummy gluey paste, the solution is not less water. The solution is time time time, and more water added slowly in stages.
3. Being whole grain, the flour you and I are working with needs more water than the other bakers who are using endosperm-only durum. Our hydration will need to be in the 85% to 90% range.
Side note: semolina and semolina rimacinata does not behave like this, so the "culprit" must be the bran. The bran is somehow interfering with how our flour hydrates, so we need to figure out a different approach to how we hydrate our whole grain durum.
Note: bran absorbs water differently (different speed and different amount) than endosperm. You already know this: WW just hydrates and handles differently than white endosperm-only flour.
again, Note: Durum bran is going to behave differently than red/white wheat bran. If durum is not red/white wheat, then durum bran is likewise not red/white bran. How is it different? Let's abandon assumptions and explore!
(the first assumption to abandon is that what was retained in the seive is bran. So to simplify, do.... not..... sift.)
I think I figured this out with Kamut which is closer to durum than to red/white wheat. Kamut is also vitreous / glassy like durum.
I have made home-milled Kamut, but not durum.
And what made my home milled Kamut "bakeable" for me was.... soak time.
Your stone ground whole grain durum will have larger particles than my roller milled whole grain durum. So... that initial wait time after you add the first water at 77% could be as high as 8 hours as opposed to my 3 hours for roller milled whole grain durum.
What I suggest is ___establish a hydration baselne__, like how I discovered my 77%.
Take 4 bowls. Put 100 grams unsifted durum, and 2 grams salt, in each. Hydrate each one differently: 70%, 75%, 80%, 85%.
Cover and let stand 8 to 12 hours. Then... knead each sample... and see which ones are now workable dough, and which ones are still gluey paste.
The highest hydration that is workable dough is your first iteration or approximation of a baseline.
Forget, toss out, the higher hydration samples that are still gluey after the 8 hours. In my experience, something happens, where you can't "undo" the gluey nature. Again, the notion that we can "correct" the glue situation by adding flour ..... comes from our experience with red/white wheat, and durum is just not going to act like red/white wheat. (Maybe there is a "fix", but I haven't discovered it yet.)
Now... Add 4 grams water to the lower hydration samples that became workable dough.
The samples will all likely turn to gluey paste, which happens to me. but as before... give them time. Say 45 minutes.
The question now becomes.... how high hydration can you go and still have the "paste" revert to "workable dough" after giving it time to absorb?
so.... 77%, wait 8 hours, add 4%, wait 45 minutes, add 4%, wait 45 minutes.
But now, don't throw out anything that is still paste after 45 minutes. Just set it aside and see if it just needs more time. Your magic wait period might be 60 minutes.
My answer for roller milled flour is 30 min wait times and a max 92%. But I can still get a good loaf at 89%, which is what I shoot for now.
Yours could be more or less, as your durum grain might have more or less native moisture. And your time-to-absorb will be be longer than mine due to home-milling likely creating larger particles than roller-milled.