MATRIX ACIDLUNG OF SANDSTONES
[Al(OH)J in the gibbsiteform canprecipitateupon spend-
ing of the acid. AIFs is especiallylikely to precipitatewhen
the HF/HCl ratio and HF concentrationare high (e.g.,
1:4 or higher) (Walsh et al., 1982). These precipitates
form according to the reactions:
Al3 + 3F- + AlF3,
A13+ + 3OH- + Al(OH)3,
with
K = I()--32.5
s quartz surfaceby electrostaticinteractions, for its point
Ferric Hydroxide Or Carbonates
Ferrous or ferric ions in the acidizing mixture can come
from pipe rust incorporated by shear effect of flow or
releasedfrom iron-bearing minerals (chlorite, hematite,
pyrite, and glauconite). The usual ferrous/ferric ion ra-
tio encounteredin acidicmixturesis 5-10:1 (Crowe, 1985).
This ratio takes into account.only the iron from iron-
bearingminerals.Most iron is dissolvedinto the acid mix-
turesby simplecontactwith the steelfrom tanksandpipes.
This dissolved iron occurs mainly as ferric ions (rust),
which are potentially damaging.
Someof the ferric ions from the pipe rust can be elimi-
nated during pumping through the reaction with tubing
metallic iron:
2Fe3’ + FecsOlid) + 3Fez+.
The sametype of reduction can occur naturally in sour
wells, as a result of the action of hydrogen sulfide:
2Fe3’ + H$ + 2HZ0 + S(s)+ 2Fe” + 2H30’.
In additionto the depositionof elementalsulfur, insolu-
ble ferroussulfidecaneasilybeprecipitatedat a pH greater
than 1.9 by the reaction with excessacidic hydrogen sul-
fide (H$YHS- with Ka = 10m7):
Fe” -k H$ + HZ0 + FeS + HS- + H30’.
Ferrous‘ions can hardly lead to precipitated ferrous
hydroxide in the near-wellborearea, sinceit forms when
the pH is higher than 6. Ferric iron is normally more
problematicas ferric hydroxide precipitatesabovepH =
2. However, in the presenceof fluoride anions,ferric ions
are complexedand Shaughnessyand Kunze (1981) have
shown that precipitation of ferric hydroxide doesnot oc-
cur until there is a pH of 5 to 6:
Fe3’ + 3OH- + Fe(OH)3,
which precipitates at pH > 2.2 as KS = 10m3’;
FeFzmn(complexes) + 3OH- + Fe(OH)3,
which precipitates at pH > 5.
This mechanismof forming iron fluorine complexesap-
plies only to relatively clean sandstones.In the presence
of clays, the dissolvedaluminum ions have a greater af-
finity for fluorine than iron does.Thereforethe iron fluo-
rine complexeswill not form and iron hydroxide will still
precipitate at pH > 2.2.
The natureof the precipitate(crystallineor amorphous)
varies as a function of the anions present (Smith et al.,
1969). Ferric hydroxide can be strongly bound to the
of isoelectric chargelies abovepH 7. In the presenceof
excesscalcite, the dissolvedcarbon dioxide can also lead
to precipitation of insoluble ferric carbonates(siderite or
ankerite).
Iron-bearingprecipitatescanbe preventedby maintain-
ing a low pH in the damagedzone, by reducing anaero-
bic conditions, and by using clean equipment. If a large
quantity of iron is present, adding suitable complexing
agents in acid can prevent ferric hydroxide formation
(Crowe, 1984b). Usual complexing agentsare:
. acetic acid (up to 150’ F),
l citric acid (up to 275 ’ F),
l tetrasodium salt of ethylenediaminetetraceticacid
(EDTA) or nitrilotriacetic acid (NTA) (up to 350 ’ F);
more stablecomplexesareformed with EDTA andiron
than with EDTA and aluminum.
The possibleprecipitation of organic acid calcium salts
often usedat high temperaturesmust be consideredwhen
using theseagents.The addition of strongreducingagents
is also a good means of preventing ferric hydroxide
precipitation (Crowe, 1985).
l EDTA plays this role above 250’ F.
l Erythorbic acid works up to 400 oF with an efficiency
of 7 to 10 times that of citric acid.
14-5.3 Kinetics: Factors Affecting Reaction Rates
This subsection summarizes qualitatively the results
describedin detail in Chapter13. Sincetheoreticalaspects
were covered in Chapter 13, only practical implications
are discussedhere.
Kinetically controlled reactions (surface reaction-
limited) are effective during the acidization process of .
sandstones,and factors affecting reaction rates are dis-
cussed to complete previous thermodynamic
considerations.
Hydrofluoric Acid Concentration
Dissolution reaction rates are proportional to the
hydrofluoric acid concentration(Fogler et al., 1976;Kline
14-11
RESERVOIR STIMJLATION
et al., 1981) for most sandstoneminerals, except smec-
tite. This explains why formations with low competence
(weak cementation, potentially mobile fine particles)
shouldbe treatedwith a reducedstrengthmud acid (1.5%
HF) to avoid crumbling, especially at bottomhole tem-
peraturesgreater than 200 ’ F. Fluoboric acid performs
similarly becauseof the very low concentrationof HF
present at any time.
Hydrochlotic Acid Concentration
Dissolution reaction rates generally increasein a more
acidic medium, as the leachingof constitutivesurfaceca-
tions involves their replacementby protons, but the de-
pendency on HCl concentration is not straightforward
(Gdanski, 1986). The principal role of hydrochloric acid
is to prevent secondaryprecipitations by maintaining a
low pH. The other main effect of hydrochloric acid is to
catalyzethe attack of sandstonemineralsby hydrofluoric
acid. The mechanismand degreeof catalysisdependon
the type of mineral, as shown below.
l For instance,the reaction rate measuredat 35 ’ ‘C for
pure quartzhasthe following expression(Fogler et al.,
1976):
= 9.2 x lo-’ (1 -I- 0.8 [H+]) [HF]
c?quartz
(moles quartz/cm2/s). (14-l)
l , In the caseof a feldspar having the Na - 0.72K -
0.08Ca - 0.2Al - 1.2Si - 2.80g overall formula,
the following expressionhasbeen determined(at 25 o
C under 275 kPa pressure)asthe reactionrate (Fogler
et al., 1976):
qfeldspar = 1.3 x lo-’ (1 + 0.4 [H+]) [HF]
(moles feldspar/cmVs). (14-2)
An elementalmechanismhasbeenproposedto explain
the previous variation. It involves the adsorptionof pro-
tons on the surfacethat weakensthe siloxane bandings,
followed by the reaction of hydrofluoric acid molecules
that createsunstablesilicon-fluorine bondsat the surface,
according to the scheme:
H the previous feldspar, it is about 8 kcal/mole (Fogler et
I al., 1976).
-x -0 -Si- +Ht - -X- 0 . ..Si’ - + HF -
-X -I- FSi- + H20
where
X = Al or Si.
14-12
This is the acid (proton) catalysismechanismproposed
by Kline (1980) for feldspar.
l Dissolutionreactionis a first-orderreactionwith respect
to HF concentrationfor most aluminosilicate miner-
als. Nevertheless, dissolution kinetics is better
representedby a Langmuir-Hinshelwood type law in
the caseof sodiummontmorillonite (Kline and Fogler,
1981):
KKadsP-W
R= (14-3)
1 + KadsD-1 ’
in which KOdsis the equilibrium constant of the ex-
othermic adsorptionof HF moleculesat surfacereac-
tive sites. This adsorptionconstantis independentof
the total acidity, while K increaseswith proton con-
centration (acid catalysis). KadJis especiallyhigh for
a mineral having a high cationexchangecapacity,such
assodiummontmorillonite. For mostother clay miner-
als the value of this adsorptionconstantis small.There-
fore, when 1 B KadJHF] the expressioncan be sim-
plified to the experimentallydeterminedfirst-order ki-
neticslaw. An elementalmechanismdifferent from that
mentionedfor feldsparscanbe proposedto explain the
kinetics and to take into account solely the HF
adsorption:
F
I
~0 OH HO H OH HO OH
I I Ill I I
-X-O-X- +HF+ -X-O-X- Y=-X-F+HO-X-
where
X = Al or Si.
Temperature
Dissolutionof mineralsis a thermally activatedphenome-
non; thus, the rates increasedrastically as a function of
temperature(approximatelymultiplied by two for quartz
for a 25 ‘C increment), and the penetrationdepthsof live
acid diminish accordingly. In the caseof quartz, the acti-
vation energy is about 5.2 kcal/mole, and in the caseof
Fig. 14-5showsthe variationof the reactionrateof mud
acid with vitreous silica (more reactive than quartz) as
a function of both HF concentration and temperature
(Smith andHendrickson,1965).Aluminum andiron solu-
biIities also increaseslightly with a rise in temperature.
 MATRIX AUDIZING OF SANDSTONES

Pressure
A pressure increase speedsup the overall dissolution reac-
tion slightly, as dissolved silicon tetrafluoride can be trans-
formed partially into an acidic speqies (HZSiFh) and can
quickly initiate further reactions. For quartz, a rise of 24 %
in reaction rate was noticed between the two extreme con-
ditions (Smith et al., 1965).
In a radial injection situation, the mineral pore space
texture that determines flow partition around the wellbore
(most live acid flows through the large pores) is also a
Reaction Time, 60 min relevant parameter; clay clasts can be bypassedby the acid
flow (Williams, 1975).
‘0 1 2 3 4 5 6 7 8 9 10
HF (O/o)
14-5.4 HF Acid Reaction Modeling
The parameters that affect the reaction rate of HF acid
Figure 14~5-Reaction rate of HCI-HF on silicate glass. (Af- on sandstoneminerals are incorporated into a model that
ter Smith and Hendrickson, 1965.) predicts the evolution of formation parameters when acid
is injected.
Mineralogical CompositionAnd AccessibleSurfaceArea
In terms of surface reaction rates, sandstonesare typi-
The total specific surface area of sandstone rocks is rela- cally considered a two-component system:
tively high, and mud-acid spendingis determinedprimarily
l a slow reacting pseudocomponent,comprising the crys-
by this parameter becauseof the heterogeneous nature of
talline quartz fraction,
the dissolution reaction. However, if the contribution of
each mineral to the total accessible surface area is consi- l a fast reacting pseudocomponent, comprising all other
dered, great discrepancies between the reaction rates of species; i.e, clays, feldspars, and poorly crystallized
pure phases can be predicted and observed as shown in silica.
Table 14-5. For both pseudocomponents,the overall kinetics, which
includes the diffusion of HF-reactant species to the sur-
face; the surface reactions; and the diffusion of reacted
I Minerals I Wecific Area I products back to the bulk solution; is governed by the sur-
Quartz CO.1 cmYg face reactions because they are the slowest step. There-
Feldspars Few mz/g fore, matrix acidizing (HF) of sandstonesis often called
Kaolinite 15-30 mYg
“surface reaction-limited. ”
Mite 113 mz/g
Smectite 82 m2lg This is the major difference with matrix acidizing of
limestones, where the process is ‘ ‘diffusion-controlled. ’ ’
Table 14-5-Relative specific areas of sandstone minerals. In sandstones, the increase in permeability results from
damage removal and is correlated with a small increase
Clays react much faster than feldspars, and feldspars in rock porosity. Quartz reacts very slowly with
react much faster than quartz especially in the presence hydrofluoric acid; reactions with most aluminosilicates
of high proton (H+) concentrations. Thus, most of the provoke a rapid spending of the acid. A pseudostation-
quartz matrix can be considered as inert with respect to ary state reflects the much faster variation in species con-
the dissolution reaction (about 95 %), and the mineralog- centration (chemical modifications) than the one within
ical nature of accessible rock components determines the the rock porosity (resulting in physical modification). The
overall reaction rate. Calcite reacts at the highest rate of HF acid progresses and dissolves homogeneously every
all sandstone minerals with HF microchanneling, but in pore (and never forms conductive channelsor wormholes).
this case the mechanism of attack is not comparable be- The flow is stable and sharp fronts are formed to illus-
cause protons coming from either hydrochloric or trate the dissolution of different mineral species as acid
hydrofluoric acid can provoke the dissolution. injection progresses radially (McCune et al., 1975).

14-13
RESERVOIR STIMULATION
Severalauthorshavetried to model sucha process.Taha
et al. (1986) use the reaction model developedby Fogler
and various coworkers (see particularly Hekim et al.,
1982). Such a simplified, two-pseudocomponentmodel
and macroscopicdescription can be usedbecauseFogler
has shownthat the order of reactionof HF acid with each
pseudocomponentis equalto unity relative to the concen-
trations of HF and of the pseudocomponent.The flow is
consideredstable.
In such a model, the mineral dissolution fronts can be
computed and the concentrationof remaining clays (or
fast reactingmaterials)canbe calculated.Then, the perme-
ability increasecanbe estimatedfrom the changein porosi-
ty (or amountof material dissolved). The velocity of the
mineral dissolution front dependson the “acid capacity
number,’ ’ which is a function of the volume of clays (or
fast reacting dissolvable material) and of the acid
concentration.
The acid concentration (or spending) front can be
modelled similarly. The thickness of the front depends
on the Damkohlernumber,which is a functionof the reac-
tion rate andthe acid velocity. Suchsimulationsshowwhy
the HF acid does not penetratedeeply into the reservoir
before spendingunlessunrealistically large volumes are
used.(Thesewould almostdissolveeverythingaroundthe
wellbore and thus leave the reacted formation totally
unconsolidated.)
14-6 OTHER ACIDIZING FORMULATIONS
Problemsrelatedto the useof mud acid to removedamage
in sandstoneformations include:
l Rapid spendingprovides only a short penetration, es-
pecially at high temperatures(maximum depth about
30 cm).
l Fines, composedof either mostly quartzor mostly clay
minerals,canbe generatedduring the acid reactionand
can migrate with fluid flow. Destabilization of fines
can leadto a quick production decline after treatment.
Gravel-packedgaswells canexhibit a 50% productivity
reduction.
l The high dissolving power of mud acid destroysrock
integrity at the formation face.
New sandstoneacidizing systemsare designedto al-
leviate these shortcomings.
14-6.1 Fluoboric Acid
Fluoboric acid has been recommendedby Thomas and
Crowe (1981) as an alternative to mud acids. It doesnot
contain large amountsof hydrofluoric acid at any given
14-14
time and thus has a lower reactivity. However, it gener-
ates more HF, as HF is consumed,by its own hydroly-
sis. Therefore, its total dissolving power is comparable
to a mud-acid solution. Fluoboric acid solutionsare used
as a preflush before treating formations sensitiveto mud
acid; this avoids fine destabilizationand subsequentpore
clogging. They are also used as a sole treatment to re-
move damagein a sandstonematrix with carbonatece-
ment or in fissuresthat containmany clay particles. Fluo-
boric acid is also used as an overflush after a mud-acid
treatmentthat hasremovednear-wellboredamage(up to
0.5 ft) to allow an easierpenetrationof the fluoboric acid
solution (a few ft). Fluoboric acid is also particularly
recommendedwhenthe sandstonecontainspotassicminer-
als by avoiding damagingprecipitatesand in the caseof
fines migration due to its fines stabilization properties.
In the field, it is easily preparedby mixing boric acid
(HsBO.J, ammoniumbifluoride (NHdF. HF) , and hydro-
chloric acid. Ammonium bifluoride, an acidic salt of
hydrofluoric acid, reacts first with hydrochloric acid to
generatehydrofluoric acid:
NHdF. HF + HCl - 2HF + NH&l.
Tetrafluoboric acid is formed as a reaction product of
boric acid with hydrofluoric acid, according to:
HjBOs + 3HF - HBFsOH + 2HZ0 (quickreaction),
and
HBFsOH + HF + HBFd + HZ0 (slow reaction).
Hydroxyfluoboric acid (HBF30H) probably does not
exist in aqueoussolutionsunlessit is in equilibrium with
fluoboric acid (Wamser, 1948). The last reaction above
is of an order equalto unity with respectto both HF and
H13F30H.For this reaction,equilibriumis attainedat room
temperatureafter nearly 40 minutes for a resulting 1M
HBF4 solution. Since the equilibrium constantat 25 ’ C
is KZ = 2.3 x low3 (Wamser, 1948), about 6% (molar)
HBFd is convertedinto HBFjOH at equilibrium for a 1M
HBF4 solution. These equilibrium considerationsmean
that at any giventime andplacethereis only between0.1%
and 0.2 % (weight) of free HF at ambienttemperatureand
100’ C (212’ F) respectively.
Fluoboric acid is a strong acid of strengthcomparable
to that of hydrochloric acid (Maya, 1977); thus, the fol-
lowing reaction occurs in the solution:
HBFd + HZ0 - H30+ + I3F;.
In the following text, reactionsare written using BFF
insteadof HBF+ Acid strength diminishesin this order:

fluoboric, hydroxyfluoboric [the strength of which can
be comparedto that of trichloroaceticacid (Maya, 1977)],
and boric acid (pKnsuOs = 9.2 at 25 ’ C).
The dissolvingpower of fluoboric acid resultsfrom the
generationof hydrofluoric acidthroughits hydrolysis;i.e.,
BF; + HZ0 * BF30H- + HF.
The BF30H- anionscanbe further hydrolyzed succes-
sively into BF*(OH)L, BF(OH)l andH3BOswith the cor-
relatedHF formation,but thesereactionsmustbe takeninto
accountonly when the BF30HV concentration is lower
than 3 x lOL3 at 25’ C (Wamser, 1948). In the follow-
ing text, BF30H- hydrolysis hasbeenneglectedat usual
acid concentrations.
The hydrolysis reaction kinetics of fluoborate ions is
affected by:
l the concentrationof the fluoborate ions,
l the medium acidity that has a catalyzing effect (reac-
tion is proportional to the proton concentration),
l andthe temperature,throughtheusualactivationenergy
effect.
Thus, the reaction rate, assumingthe reversereaction
is negligible, can be expressed,after Kunze et al. (1983),
as:
d VLI
q=- = K [H30+] [BFd-] (14-4)
dt
in moles/lit/min,
with
26,183
K = 1.44 x 1017exp - 1 987T
1 (14-5)
in (mol/lit) - ’ min- ‘,
and T = temperaturein K ‘.
Thus, the reaction rate is increased300-fold when the
mixture is heatedfrom 25 ’ C to 65 ’ C and is increased
12,000-fold when heatedfrom 25 o C to 105o C. Since
the hydrolysis reaction kinetics is not affected by clays,
fluoboric acid is not considereda strongly retarded acid
in normal use. In the presenceof excessbentonite, pure
0. 1M initial fluoboric acid is spentwithin 30 minutes at
150’ F (Kunze and Shaughnessy,1983).
The reaction of fluoboric acid in sandstonesinvolves
at the sametime the hydrolysis reactionof fluoboric acid,
the classicalreactionsof the generatedhydrofluoric acid
with minerals,and additionalslow reactionsrelatedto the
fluoborate ions in liquid phase.As expected,the dissolu-
MATRlX ACIDIZING OF SANDSTONES
tion reaction of clays with fluoboric acid is a first-order
reaction with respectto the fluoborate concentration, as
it is for mud acid with respectto HF concentration.
The spendingrate of fluoboric acid on glass slides at
150’ F is one-tenththat of a mud acid having the same
total HF content(ThomasandCrowe, 1981). Sinceamor-
phous silica reactsfaster than quartz, limited deconsoli-
dation of coresnear the injection face is observedduring
flow testswith fluoboric acid. This is comparedto com-
plete crushing with mud acid. (The decreasein compres-
sive strength of the treated cores is reducedby 30% to
50%.)
The unique advantageof fluoboric acid is that it pro-
vides an efficient clay/fine stabilizationthrough reactions
relatedto borateand fluoborate ions. Swellableclays are
desensitizedby fluoboric acid: a strong decreasein ca-
tion exchangecapacity (e.g., minus 93% after 18 hours
in fluoboric acid at 150’ F for a Wyoming bentonite)was
noticed ‘by Thomas and Crowe (1981).
After a fluoboric acid treatment, migrating clays and
other fines stabilize as a result of the rock’s exposureto
acid. This is the reasonfor the long shut-in time recom-
mendedin fluoboric acid treatments. During injection,
while the acid spendsnormally, coresjust treated with
fluoboric acid exhibit a normal increasein permeability.
However, there is no long-term stabilization after treat-
ment becauseonly part of the clays dissolved; the rest
did not havetime to stabilize. Additional shut-in time al-
lows this stabilization.
When treated by fluoboric acid, montmorillonite
progressively decreasesin aluminum content and then
progressively incorporatesboron atoms; silicon precipi-
tatesfrom the solution. Coresoriginally containing 30 %
silicoaluminatesat 150’ F attainedmaximum static solu-
bilities after only 24 hours in the presenceof 1M HBFd.
(4 hours in the caseof mud acid), whereasthe maximum
increasein permeability is obtainedafter only 4 hours un-
der dynamicconditions.Theseresultsprove thedissociated
effectsof mineraldissolutionby the generatedhydrofluoric
acid (essentiallykinetically controlled) and of the parti-
cle stabilization resulting from slow complex dissolu-
tiomreprecipitation mechanisms(toward thermodynam-
ic equilibrium) during the shut-in period.
Original kaolinite clay platelets-pure aluminosilicates-
that are not dissolved by fluoboric acid appear welded
together and to the quartz grains when examinedunder
a scanningelectron microscope. A type of chemical fu-
sion of any fines seemsto take place slowly onto the sili-
ca surface.The formation of borosilicate“glass” hasbeen
assumedto account for this reaction.
14-15