Parts of a Perfume Bottle: Why Two Components Marked the Same Size Still Won’t Fit

What the Parts of a Perfume Bottle Actually Are, and What Else They’re Called

A perfume bottle looks like a single object. Press the pump and it still feels like one object. In fact it is an assembly of pieces that are tooled, stocked and bought separately. And the first thing that trips people up is not the number of pieces. It is that the same piece has three or four names depending on who is speaking.

Work from the bottom up.

Those aliases matter. They are the reason a conversation stalls. Ask for a “nozzle” and a supplier may hear the actuator body, the nozzle insert inside it, or the spray assembly as a whole. Ask for a “collar” and you may be handed a decorative ring that does not crimp. The worst offender is the spray end, where three different things share one name. atomizer is the whole pump system. actuator is the button you press. nozzle is the small opening the liquid is forced through.

One more split worth clearing up early. In vintage and collector listings, “perfume bottle parts” means replaceable pieces for bottles you already own: squeeze bulbs, tassels, glass balls, lengths of tubing. This article is about the other sense: the component system of a new bottle, bought as one order. The two look identical as search phrases and have almost nothing else in common.

Where One Part Ends: the Four Kinds of Boundary on a Bottle

The four boxes an item can fall into

Counting the pieces is where published guides disagree. One guide lists five parts, another six, a third seven, and each count is describing something real. They are all describing the same object. They are drawing the boundary in different places.

There are four kinds of boundary on a perfume bottle, and a given item belongs to exactly one of them.

The test that decides which box

The test for which box an item belongs in is one question: can it be ordered on its own, and can it be replaced on its own? That single question does two jobs. It tells you what belongs on a purchase order, and it tells you what you cannot fix later.

The costs land differently in each box, and this is where counting granularity stops being academic. Published sourcing guidance for new fragrance brands puts held stock in the low hundreds of pieces. Custom caps commonly start around 1,000 to 2,000, and a custom glass mould at 3,000 to 5,000 plus tooling (Hopestone Advisory, 2026). Those are different orders of magnitude across pieces of the same product. So a low-priced bottle with a custom cap is not a low-priced order. You see why the moment you write the parts down as separate lines instead of thinking of them as “the bottle.”

Why the neck appears in none of them

The neck is the one that surprises people. It is not a part. It is a specification: a set of tolerances applied to the glass. Suppliers do not quote necks. They quote bottles, and the neck is the part of the bottle drawing that everything else has to agree with.

Crimp, Screw, Crimpless or Snap-On: the Choice You Can’t Take Back

Published guides disagree on how many closure systems exist, and here the disagreement is genuinely useful. Some describe two, some three, some four. They are not counting the same thing. At two, the question is whether the line needs a crimping machine. At four, snap-on and crimpless count as systems in their own right rather than as variations.

The four differ in ways that matter more than their names suggest.

Two practical consequences follow.

Crimpless systems do not use a separate collar the way a crimp system does. That seat already contains the collar. So “do I need to buy a collar?” has different answers depending on the system you picked, which loops straight back to the boundary question in the previous section.

And the pump is not one part either. Inside it sit an actuator, a pump chamber, a spring, at least one seal and a nozzle insert, plus the dip tube. Fine-fragrance pumps typically deliver somewhere in the range of 0.08 to 0.12 ml per stroke. That is wide enough that you should treat it as a band rather than a fixed figure. That number shapes daily use more than it looks. The dose on the spec sheet is the number of times your customer will touch the product.

The dip tube is where the arithmetic becomes visible. Cut it to the internal depth of the bottle and leave a small gap at the base. Published guidance ranges from 1–2 mm to 2–4 mm; of the two, the conservative reading is the safer one. Too long and the tube buckles against the base and restricts flow. Too short and the pump starts drawing air while product is still in the bottle. Customers report both as a defective product. Nobody reports a tube that was cut wrong.

The One Dimension Every Part Shares: Neck Finish, Tolerance, and the PO Line

A neck finish is five dimensions, not one

One surface on the whole bottle has to be agreed by four separate makers: the glass maker, the pump maker, the collar maker and the cap maker. That surface is the neck finish. It is standardised at European level by EN 14854:2021, Glass packaging — Dimensions of neck finishes for aerosol and spray glass containers. That standard covers nominal neck diameters of approximately 11, 13, 15, 17, 18 and 20 mm in both moulded and tubular glass (CEN via AFNOR, 2021). Matching ferrule dimensions are fixed by EN 14849.

The shorthand “FEA 15” hides that the finish is a cluster. It covers the sealing bead diameter at the top, the width under the flange, and the bore the dip tube passes through. It also covers the usable neck height and the position of the bead relative to the top. Each one controls a different behaviour in assembly. Two drawings that both say “FEA 15” can differ enough to fail on the filling line.

This also settles a contradiction you will run into. Some guides tell you to measure the inner diameter of the neck and read the FEA size off it. A finish is named for its nominal neck diameter in millimetres, and the bore is a separate dimension in the same cluster. Measuring the hole and reading it as the size is how buyers end up ordering pumps that will not seat.

The two tolerance bands meet here

A crimp seal uses no adhesive and cannot be adjusted after it is made. The glass and the metal have to arrive inside a narrow window at the same time.

Glass necks are typically held to about ±0.15–0.30 mm, with ±0.20 mm as the realistic production expectation, and beyond ±0.30 mm the pumps stop sealing and alcohol loss begins. The aluminium collar on the other side holds its bore to roughly +0.00 / −0.03 mm, with step concentricity around 0.03 mm TIR so the actuator does not sit at an angle. Neither side is being sloppy. Each is inside tolerance for its own drawing, and the pair still does not close.

Engineering guides that list where leaks come from put this next to undersized necks and uneven crimp pressure. The wording they use is blunt: “components from mixed sources, all labelled FEA 15, fitting the drawing but not each other.”

What to verify, and what to write on the order

You can close most of this before mass production. Four checks are cheap at sample stage and expensive later. Vacuum decay catches micro-leaks. Bubble submersion locates exactly where a leak path sits. A pressure-drop test on empty glass finds cracks and bad seams. A stability run on filled units, held and re-tested after a week or two, catches gasket swelling and pump failure. The vacuum-decay and bubble-emission methods are established packaging leak procedures (ASTM F2338 and ASTM D3078).

Then write it down. Order the pieces separately but specify them as a system.

One consequence of writing the order out this way is that the component setting your order minimum stops being invisible. It is rarely the bottle. Usually it is whichever line reaches a production minimum first, and on decorated work that is frequently the cap, the coating or the printed box rather than the glass.

What Changes at Reorder: Lock the Seam Before You Talk Price

A first order is a test. A tenth order is the business. Every reorder brings the two tolerance bands back together again. The products that suffer are the ones whose buyers chose components on unit price and discovered the cost later.

So run the purchasing sequence backwards from how it usually goes.

Lock the seam first. Put the neck finish on the drawing with its dimensions and tolerances, and cite the standard it conforms to. Treat that drawing as the fixed point everything else is measured against. Then lock the set: buy the pump, the collar and the cap from one accountable source rather than three. When three suppliers each passed their own inspection, a leak at the crimp has no owner. Then compare quotes. Per-piece savings on components are real. They just have to sit on the same page as the cost of one tolerance collision: the rework, the filling line stopped, the units already shipped.

The minimum order quantity question belongs in the same place, and it rarely lands where buyers expect. Ask for each line separately (bottle, pump, gasket, collar, cap, decoration, label, insert, box, export carton, assembly) and record the packing quantity and production minimum for each one. A low bottle minimum does not create a low finished-pack minimum if the cap, the coating or the printed box starts somewhere else entirely.

At HUA GLASS we keep standard neck openings at 13, 15 and 18 mm in both screw and crimp configurations. So the finish gets specified on the bottle drawing alongside everything else, rather than traded off against it. We also hold stock specifically to support repeat orders, which matters more at the tenth order than the first. Stock items ship in 3 days. Colour or logo modification on an existing bottle takes 10–15 days. Full custom development with a new mould runs to about 40 days end to end. Three different timelines for three different kinds of reorder. Know which one your next order is before you promise a delivery date. Customisation is quoted across shape, capacity, colour, neck opening, logo application, polishing, finishing, spray pump, cap and packaging. The neck dimension is a line you specify, not a constraint you inherit.

References