Does the slot or the element set a shallow donor? Group-V donors in silicon — materials scout,
What’s claimed and how well supported
- Labels on this page
- SPECULATIVE ×4
- Strongest label used
- SPECULATIVE
- Main caveat (quoted)
- Setting it beside Grant's open "doping as operator over typed sites" question (2026-10-14) is analysis, and the Navajo comparison is ◆ SPECULATIVE and placed last.
Counted automatically from the labels printed on this page. Counts are a reading aid, not a grade of the page.
Terms on this page (5), plain-language definitions
- Host (matrix)
- Majority element/phase whose lattice dominates the material.
- Valence (chem.)
- Bonding capacity / oxidation-related electron count of an atom in context.
- Dopant
- Intentional dilute impurity meant to tune electronic/optical/catalytic behavior without replacing the host as majority phase.
- Solute
- Intentional minority species dissolved in the host (alloying addition).
- Dislocation
- Line defect enabling plastic slip at realistic stresses; edge vs screw are the two classic geometries.
Definitions quoted from the Chem for linguists primer glossary.
Date: 2026-10-14 PT (weekday research push, evening)
Status: Materials scout. The physics is same-domain description from the sources opened below (REAL inside semiconductor physics). Setting it beside Grant's open "doping as operator over typed sites" question (2026-10-14) is analysis, and the Navajo comparison is SPECULATIVE and placed last. Not a correlate. Not a joint field.
Audience: Grant (linguistics; less chemistry). End goal: designing new metal alloys (this note is a semiconductor case, used because it is the cleanest published slot-vs-filler test found so far).
Does not: claim that a donor, a lattice site or a valley is an operator, argument or theta role; supply a chemical stand-in for a theta role; invent energies or quotes; claim the gap is closed; revise Bridge A; run the six-point ranking-edge spec.
Why this note. On 2026-10-14 Grant asked whether "verbs take theta roles, not words" has a physical-chemistry counterpart: a doping process acting like an operator whose arguments are typed sites, not particular elements. The test written down then was: find a source showing that the outcome is set by the slot type independently of which element fills it. Shallow donors in semiconductors are a textbook place where physicists state exactly this — and then say where it fails.
The case, in plain words. Put a phosphorus atom on a silicon site. Phosphorus has one more outer electron than silicon, so that extra electron is loosely held by the leftover +1 charge, like a very swollen hydrogen atom inside the crystal (the hydrogenic or effective-mass picture). In that picture the binding energy depends only on the hostMajority element/phase whose lattice dominates the material. — its dielectric constant ε and the electron's effective mass m* — and not on whether the extra electron came from P, As, Sb or Bi. The "slot" here is: a substitutional site in a given host, carrying one extra charge. The "filler" is the particular group-V element.
What is new (one sentence). The slot-only picture holds almost perfectly in single-valley hosts like GaAs, and in silicon it holds for the excited 1s states that have a node at the impurity (spread about 2 meV across P, As, Sb, Bi, all near the slot-only value of −31.2 meV), but the ground state — which piles up at the impurity core — varies from −42.71 to −71.00 meV with the element, because "atomic species differ from each other at the core region" (Saraiva et al. 2014): the host decides how much the filler's identity leaks through, and it leaks only where the state reaches the core.
1. Sources opened this run
| Source | How read | Used for |
|---|---|---|
| Saraiva, Baena, Calderón & Koiller 2014, "Theory of one and two donors in silicon," arXiv:1407.8224v1 | arXiv PDF; text layer | Effective-mass theory (EMT) is species-independent by construction; GaAs agreement; Si and Ge species dependence; valley-orbit share of binding; core-region cause; Table I (experimental A1, T2, E for P, As, Sb, Bi; experimental column cites Grimmeiss, Janzén & Larsson 1982) |
| Gamble, Jacobson, Nielsen et al. 2015, "Multivalley effective mass theory simulation of donors in silicon," arXiv:1408.3159v3 | arXiv PDF; text layer | Independent repeat of the P level means (−45.59, −33.89, −32.58 meV) |
| Zwanenburg, Dzurak, Morello et al. 2013, "Silicon quantum electronics," Rev. Mod. Phys. 85: 961 (arXiv:1206.5202) | arXiv PDF; text layer | For P, central-cell corrections lower the ground valley state by 11.7 meV and lift the sixfold degeneracy |
Verbatim passages relied on: excerpts-verbatim.md (not published here) (cut programmatically from the text layers; PDFs and text alongside).
Not opened (do not treat as read): Ramdas & Rodriguez 1981 Rep. Prog. Phys. 44: 1297 (the standard review; cited by Saraiva et al. for the Si/Ge spread); Grimmeiss, Janzén & Larsson 1982 Phys. Rev. B 25: 2627 (the measured values — numbers here are as Saraiva et al. tabulate them); Kohn & Luttinger 1955; Kohn 1957; Fetterman et al. 1971 (GaAs); Pantelides & Sah 1974; any germanium donor table; any acceptor table (B, Al, Ga, In in Si); any source on how a doping process gets the atom onto the substitutional site. An arXiv metadata query hung and was dropped; journal versions of Saraiva et al. and Gamble et al. are not cited here.
2. Sourced materials
2.1 The slot-only prediction, and where it is confirmed
- Saraiva et al. §II: EMT "involves no fitting parameters, as the binding energy is dependent on the material only through ε and m∗, leading to a value of E ∗ that is independent of the donor species." They add: "This is in fact clearly demonstrated experimentally for example in GaAs, where deviations from the hydrogenic expression are hard to detect."
- Their Fig. 1 plots measured vs EMT binding energies for single-valley hosts (the plotted labels include GaAs with S, Se, C, Si, Ge and InSb with Te): "The agreement is very good mainly for the less bound donors."
- So in a single-valley host, the slot sets the answer and the fillers are interchangeable, as the source states it.
2.2 Silicon: the same slot, a filler-dependent ground state
- Same paragraph: "In contrast, the binding energy of donors in Si and Ge is strongly dependent on the donor species." Fig. 1 caption: Si ground states "present a large variation (∼ 50%) as a function of the donor species."
- The slot-only (single-valley Kohn–Luttinger) value in Si: "EKL = −31.2 meV."
- Table I, experimental, as tabulated by Saraiva et al.:
| Donor | Ground A1 (meV) | T2 (meV) | E (meV) |
|---|---|---|---|
| P | −45.58 | −33.90 | −32.60 |
| As | −53.77 | −32.65 | −31.40 |
| Sb | −42.71 | −33.00 | −30.60 |
| Bi | −71.00 | −31.92 | N/A |
My arithmetic on those numbers (not the authors'): the A1 values span 28.29 meV (−42.71 to −71.00) and lie 11.5 to 39.8 meV below EKL; the T2 values span 1.98 meV and the E values 2.00 meV (Bi E missing), all within about 2.7 meV of EKL. Gamble et al. use (−45.59, −33.89, −32.58) meV for P, matching to 0.01 meV. Zwanenburg et al.'s "lowered by an energy 11.7 meV" for P matches A1 minus T2 (45.58 − 33.90 = 11.68). - Table I caption: "the experimental and calculated values for the T2 and E energies do not depend strongly on the donor species and are very similar to the single valley result EKL, in contrast to the value of the ground state A1 energy."
2.3 Why: the filler shows only at the core
- §III D: "The T2 and E states are much less sensitive to the effect of Vcc since both have nodes at the impurity site." (Vcc = the central-cell correction to the potential.)
- §III B: the screened point-charge picture "is realistic only at large distances from the impurity site. Close to the nucleus, the dielectric response becomes inhomogeneous. Furthermore, the valenceBonding capacity / oxidation-related electron count of an atom in context. core shells for the impurity are different from the Si atoms, presenting a strong departure from the excess proton picture. Atomic species differ from each other at the core region, leading to the strong donor species dependence of the binding energy in silicon in contrast with other semiconductors."
- §II end: the coupling among Si's six conduction-band valleys, mediated by the impurity, "accounts for over 30% of the binding energy in silicon donors."
- In the modelling the authors describe, the element enters as a fitted core parameter (r_cc) chosen per species to hit the measured A1; everything else is shared.
2.4 Summary of the materials picture
- A shallow donor's long-range behaviour is set by the slot (host ε and m*, one extra charge on a substitutional site).
- Filler identity enters only through the short-range core, and only for states that have weight at the core.
- Whether that core detail matters much is decided by the host: barely in GaAs (single valley), strongly in Si and Ge (multivalley, the core couples the valleys).
3. Analysis beside the operator question (analysis, not a finding)
Grant's test: does the outcome depend on the slot type independently of which element fills it?
- Partly yes. In single-valley hosts, as the source states, yes. In silicon, yes for the excited 1s states (about 2 meV spread over four elements, near the slot-only value) and for the long-range tail of every state.
- Partly no. The silicon ground state — the one that matters most at low temperature — carries the element's fingerprint (28 meV spread). The physics literature handles this by keeping the slot model and adding one per-element correction at the core, not by switching to an element-by-element model.
- What this does not test. It is about the state of a dopantIntentional dilute impurity meant to tune electronic/optical/catalytic behavior without replacing the host as majority phase. already sitting in a slot. It says nothing yet about a doping process choosing the slot (substitutional vs interstitial, clustering), which is element- and process-dependent and was not opened. So it supports "typed slot with a filler-local correction" for the result, not "the process selects slot types" for the operation.
- Ranking-edge spec: not run.
SPECULATIVE Cross-domain section. Open after the materials.
4. Navajo comparison — SPECULATIVE (after the materials)
Only as a shape comparison, not a correlate: in the Navajo work (Navajo Decoded; Willie 1991, cited at p. 52 on the Navajo site; text layer of the scan: "The noun by itself cannot be an argument it must be licensed by a pronoun" as the scan's text layer reads, the dash dropped), the role-bearing pronoun slot sets the clause's argument structure and a noun is a linked add-on. The donor case has the same kind of division of labour — a shared slot template plus a filler-specific correction that shows up in one place. In Navajo the obvious place the filler's identity shows through is yi-/bi- choice, which reads the noun's rank (animacy and the proto-agent-like properties on the Navajo site). Whether that is more than a resemblance is open. No chemical stand-in for a theta role; Bridge A unrevised; not a joint field.
5. Open questions
- Germanium donors and silicon acceptors (B, Al, Ga, In): does the same "excited states shared, ground state element-specific" split hold? (Ramdas & Rodriguez 1981 not opened.)
- The process side of Grant's question: is substitutional vs interstitial site choice for a dopant described as set by the site type or by the element? (Not opened.)
- Metal-alloy counterpart: is there a metal-host property with the same split — long-range set by the slot, short-range by the element (for example soluteIntentional minority species dissolved in the host (alloying addition). strengthening by size and modulus misfit vs chemical bonding terms)? Candidate for a later push. → Taken up 2026-10-15 evening in Solute strengthening: host part vs element part: yes — in Mg (Yasi et al. 2010) a long-range size term set by the host's dislocationLine defect enabling plastic slip at realistic stresses; edge vs screw are the two classic geometries. field plus an element-specific core ("chemical") term, size dominant; STRONG.
Corpus note: si-donor-slot-vs-element-v0.md. The metallurgy is same-domain. Setting it beside the Navajo causatives is analysis at event-structure grain, not a correlate. No chemical stand-in for a theta role. Grant’s phrase↔molecule frame is speculative. Not Bridge A. Not a joint field.