PDA is same like UV in some cases UV more sensitive than PDA.
RI - is good for analyze carbohydrates, triglycerides, organic acids, and pharmaceutical excipients. RID is the detector which also can be choice for polymer characterization using gel permeation chromatography (GPC) and size exclusion chromatography (SEC).
ELSD is better for non-UV absorbing compounds, its a primary choice since the principle of detection does not rely on the optical properties of the solute and its able to detect all compound which are less volatile than the mobile phase.
Based what exact components you want to analyse you can chose one of above options.
PDA (diode array detector). You get some idwntyfication based on UV spectrum. Second pne in series like RI or ELSD get be useful for estimation of its concentration level - if you don't have standards.
PDA is same like UV in some cases UV more sensitive than PDA.
RI - is good for analyze carbohydrates, triglycerides, organic acids, and pharmaceutical excipients. RID is the detector which also can be choice for polymer characterization using gel permeation chromatography (GPC) and size exclusion chromatography (SEC).
ELSD is better for non-UV absorbing compounds, its a primary choice since the principle of detection does not rely on the optical properties of the solute and its able to detect all compound which are less volatile than the mobile phase.
Based what exact components you want to analyse you can chose one of above options.
I personally like ELSD to complement UV detection. You need to keep in mind that this is a destructive technique though, so you have to dedicate some of your flow to this detector and you will need a flow splitter to do this.
I would always recommend PDA first, for its general utility, unless the class of natural products, doesn't have good UV absorbance (such as carbohydrates, most terpenes, lipids, etc.). Next choice is ELSD or MS, depending on whether quantitation or characterization of unknowns is higher priority.
If you already have a single wavelength UV/VIS, then just buy a scanning DAD (PDA). That will be far more useful. A big NO on the ELSD. You have not specified the sample and with an ELSD you would need to split the flow heavily with your planned PREPARATIVE HPLC application (sensitivity may be terrible and linearity even worse?). ELSD is very flow sensitive so even with a splitter, the baseline may be unstable (I have used ELSD with analytical and Prep HPLC systems). I have run thousands of samples with ELSD, CAD, EC, RI, UV/VIS, MS and plenty of other detectors. *A few things that the ELSD sales people may forget to mention to you: ELSD also requires a lot of regular internal cleaning, a great deal of expertise to use (Not for novice users), has non-linear output across most ranges, not a "universal" detector (LOL !), it will not see most volatile samples and uses A LOT of ultra-pure nitrogen gas (you will need a source of gas to run it, similar to an LC-MS (ESI)).
Diode array (DAD) is the way to go, esp when you have not identified the sample type yet. DAD, sometimes called PDA (PDA = DAD = UV/VIS), has a wide range, you can adjust the bandwidth to compensate for low or high signals and prep flow cells with narrow or even variable path length flow cells are readily available. It will also allow you to run in isocratic mode (common with prep) or gradient mode. If possible, get one with a variable path-length flow cell (or very narrow path length to deal with the high concentration of samples it is expected to 'see').
RID is extremely temp sensitive, but in isocratic mode can be useful with Prep applications (will detect many compounds not seen by UV), but... it is of course limited in many ways too (super temperature sensitive). Nothing is perfect, but for general use, a diode array detector is the best choice.
Depends if vast majority of your compounds absorbs in UV. But I would go for both PDA and ELSD, so you will see "almost" everything. Yes ELSD is destructive detector but its so sensitive that you can dedicate something like 0.1-1% of your column flow into ELSD and collect rest 99%+ so you will lose virtually nothing.
Majority of natural compounds are detected on UV/PDA except few ones on ELSD as most of these components contains chromophoric groups, which absorbs in UV-VIS regions.If you are targeting natural compounds such as -Isoflavones,anthocyanins and polyphenols then u can try with PDA as u can scan the entire UV-VIS region to get the optimum wavelength which will save your time to optimize the detector wavelength.